CN116151878A

CN116151878A - Recommendation processing method, device, equipment, program product and storage medium

Info

Publication number: CN116151878A
Application number: CN202111375857.8A
Authority: CN
Inventors: 秦嘉; 刘晓煜; 陈伟民
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-11-19
Filing date: 2021-11-19
Publication date: 2023-05-23

Abstract

The application provides a recommendation processing method, a recommendation processing device, electronic equipment, a computer program product and a computer readable storage medium; the method comprises the following steps: acquiring first attribute information of an object to be served, and determining individual scores of a plurality of indexes of a recommended service object; determining the information entropy of each index according to the independent scores of a plurality of indexes of a plurality of recommended service objects, and determining the weight of each index according to the information entropy of each index and the first attribute information; for each recommended service object, weighting the individual scores of a plurality of indexes of the recommended service object according to the weight of each index to obtain the comprehensive score of the recommended service object; and according to the comprehensive score of each recommended service object, performing first descending order sorting on the plurality of recommended service objects, and executing recommendation operation on the objects to be served. By the method and the device, the recommendation accuracy of the service object can be improved.

Description

Recommendation processing method, device, equipment, program product and storage medium

Technical Field

The present application relates to internet technology, and in particular, to a recommendation processing method, apparatus, electronic device, computer program product, and computer readable storage medium.

Background

With the development of the internet, service forms become various, and a selection range of a recommended service object that provides a service to an object to be served becomes larger, for example, for an advertisement delivery service, an advertiser as an object to be served needs to select from thousands of advertisement service agents (recommended service objects), the difficulty of selection is large, and service selection efficiency of the recommended service object is suppressed.

In the related art, the recommended service objects are ranked according to a certain index and recommended to the advertiser, so that the advertiser can select according to the ranking result, and the selection range of the advertiser is limited and the recommended service objects matched with the preferences of the advertiser are difficult to obtain due to the single ranking mode.

Disclosure of Invention

The embodiment of the application provides a recommendation processing method, a recommendation processing device, electronic equipment, a computer program product and a computer readable storage medium, which can improve the recommendation accuracy of a service object.

The technical scheme of the embodiment of the application is realized as follows:

the embodiment of the application provides a recommendation processing method, which comprises the following steps:

acquiring first attribute information of an object to be served, wherein the first attribute information is used for representing preference degree of the object to be served for each index of a recommended service object;

Determining a separate score for a plurality of the metrics for each of the recommended service objects;

determining information entropy of each index according to independent scores of a plurality of indexes of a plurality of recommended service objects, and determining weight of each index according to the information entropy of each index and the first attribute information;

for each recommended service object, weighting the individual scores of a plurality of indexes of the recommended service object according to the weight of each index to obtain the comprehensive score of the recommended service object;

and according to the comprehensive score of each recommended service object, performing first descending order sorting on the plurality of recommended service objects, and executing recommendation operation on the objects to be served according to the first descending order sorting result.

The embodiment of the application provides a recommendation processing device, which comprises:

the system comprises an acquisition module, a recommendation module and a storage module, wherein the acquisition module is used for acquiring first attribute information of an object to be served, wherein the first attribute information is used for representing the preference degree of the object to be served for each index of a recommended service object;

an index module for determining individual scores of a plurality of said indexes for each said recommended service object;

The index module is further configured to determine an information entropy of each index according to the individual scores of the plurality of indexes of the plurality of recommended service objects, and determine a weight of each index according to the information entropy of each index and the first attribute information;

the index module is further configured to weight, for each recommended service object, individual scores of a plurality of indexes of the recommended service object according to a weight of each index, so as to obtain a comprehensive score of the recommended service object;

and the sorting module is used for sorting the plurality of recommended service objects in a first descending order according to the comprehensive score of each recommended service object, and executing the recommending operation to the object to be served according to the first descending order sorting result.

In the above scheme, the obtaining module is further configured to obtain second attribute information of the object to be served before determining separate scores of the multiple indexes of each recommended service object, and perform at least one first screening process on multiple candidate recommended service objects according to the second attribute information, so as to obtain a first screening result; acquiring third attribute information of the object to be served, and inquiring candidate recommended service objects matched with the third attribute information of the object to be served from the first screening result to form a target screening result set, wherein the third attribute information is used for representing type information of the object to be served; and determining at least one recommended service object matched with the object to be served according to the target screening result set.

In the above solution, when the second attribute information includes industry information of the object to be served, the obtaining module is further configured to: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining total incomes of the candidate recommended service objects, incomes of the candidate recommended service objects corresponding to the industry information and incomes of the candidate recommended service objects corresponding to the industry information; determining a first industry index that is inversely related to the total revenue of the candidate recommended service object and is positively related to the revenue of the industry information corresponding to the candidate recommended service object; determining a second industry index that is inversely related to revenue corresponding to the industry information and that is positively related to revenue corresponding to the industry information for the candidate recommended service object; and when the first industry index is larger than a first threshold value and the second industry index is larger than a second threshold value, reserving the candidate recommended service object.

In the above solution, when the second attribute information includes at least one level information corresponding to the object to be served, the obtaining module is further configured to: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining, for each of the ranking information, the incomes of the candidate recommended service object corresponding to the industry information and the ranking information, the incomes of the candidate recommended service object corresponding to the industry information; for each of the ranking information, determining a positive correlation of revenue for the industry information and the ranking information corresponding to the candidate recommended service object, a positive correlation of revenue for the industry information, a negative correlation of revenue for the industry information corresponding to the candidate recommended service object, and a ranking index of negative correlation of revenue for the industry information and the ranking information; and when the grade index obtained for any one grade information meets the grade preference condition, carrying out reservation processing on the candidate recommended service object.

In the above solution, the obtaining module is further configured to: when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining the candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining that a direct recommendation condition is met; and when the direct recommendation condition is met, according to the candidate recommended service objects in the target screening result set, executing recommendation operation to the to-be-served objects.

In the above solution, the obtaining module is further configured to: when the type information characterization is used for sequencing the matched candidate recommended service objects in an independent index sequencing mode, determining the candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization sorts the matched candidate recommended service objects in an independent index sorting mode, determining that an independent index sorting condition is met; and when the independent index sorting condition is met, determining independent scores of indexes of the plurality of recommended service objects corresponding to the type information, sorting the plurality of recommended service objects in a second descending order according to the independent scores, and executing recommendation operation to the objects to be served according to a second descending order sorting result.

In the above solution, when the type information indicates that the matching candidate recommended service objects are to be ranked in a non-independent index ranking manner, and the number of the matching candidate recommended service objects exceeds a number threshold, the obtaining module is further configured to: and obtaining fourth attribute information of the object to be served, and performing at least one second screening treatment on the target screening result set according to the fourth attribute information to obtain at least one recommended service object matched with the object to be served.

In the above solution, the fourth attribute information includes at least one flow channel information of the object to be served, and the obtaining module is further configured to: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: determining, for each of the traffic channel information, the incomes of the candidate recommended service object corresponding to the industry information and the traffic channel information, the incomes of the candidate recommended service object corresponding to the industry information; for each piece of traffic channel information, determining a traffic index positively correlated with the incomes of the industry information and the traffic channel information corresponding to the candidate recommended service object, positively correlated with the incomes of the industry information, negatively correlated with the incomes of the industry information corresponding to the candidate recommended service object, and negatively correlated with the incomes of the industry information and the traffic channel information; and when the flow index obtained for each piece of flow channel information meets the channel preference condition, carrying out reservation processing on the candidate recommended service object.

In the above solution, the obtaining module is further configured to: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: and when the candidate recommended service object accords with each piece of capability requirement information, carrying out reservation processing on the candidate recommended service object.

In the above solution, the index module is further configured to: determining an index value of each of the indexes of each of the recommended service objects; performing standardization processing on the index value of each index of each recommended service object to obtain a standardized index value; and summing the standardized index values of the indexes corresponding to the plurality of recommended service objects according to each index to obtain a first summation result, and determining the proportion of the standardized index value of the index corresponding to each recommended service object to occupy the first summation result as an independent score of the index corresponding to each recommended service object.

In the above solution, the index module is further configured to: for each of the indices, the following processing is performed: carrying out logarithmic processing on the individual scores of the indexes corresponding to each recommended service object, and multiplying the logarithmic processing result with the individual scores of the indexes corresponding to the recommended service objects; carrying out summation processing on the multiplication processing results of the plurality of recommended service objects to obtain a second summation result; an information entropy of the indicator that is inversely related to the second summation result is determined.

In the above solution, the index module is further configured to: sorting the indexes in a descending order according to the preference degree of each index; determining at least one index which is ranked first as a preference index, and determining the rest indexes as non-preference indexes; the following processing is performed for each of the preference indicators: determining an optimized information entropy positively correlated with the information entropy of the preference index and the first minimum information entropy and negatively correlated with the second minimum information entropy; wherein the first minimum information entropy is the minimum information entropy in the information entropy of a plurality of indexes, and the second minimum information entropy is the minimum information entropy in the information entropy of at least one preference index; the following processing is performed for each of the non-preference indicators: determining the information entropy of the non-preference index as the optimized information entropy corresponding to the non-preference index; summing the optimized information entropy of the indexes to obtain information entropy sum; the following processing is performed for each of the indices: a weight of the index that is inversely related to an optimized information entropy of the index and positively related to the information entropy sum is determined.

In the above solution, the sorting module is further configured to: acquiring industry information and grade information of the object to be served; inquiring a plurality of historical objects to be served corresponding to the industry information and the grade information; determining the number of historical recommended service objects of each historical object to be serviced; averaging the number of the historical recommended service objects of the historical objects to be serviced to obtain an average number; acquiring the recommended service objects which are ranked ahead and meet the average number in the first descending ranking result; and according to the recommended service objects which meet the average number, executing recommendation operation to the objects to be served.

An embodiment of the present application provides an electronic device, including:

a memory for storing executable instructions;

and the processor is used for realizing the recommended processing method provided by the embodiment of the application when executing the executable instructions stored in the memory.

The embodiment of the application provides a computer readable storage medium, which stores executable instructions for implementing the recommended processing method provided by the embodiment of the application when being executed by a processor.

Embodiments of the present application provide a computer program product, including a computer program or instructions, which when executed by a processor, implement the recommended processing method provided in the embodiments of the present application.

The embodiment of the application has the following beneficial effects:

the method and the device can comprehensively consider the personalized requirements of the index preference of the object to be served and the individual scores of the objective indexes of the recommended service object, so that the recommended service object recommended to the object to be served can be guaranteed to be well matched with the actual situation and the appeal of the object to be served, and the recommendation accuracy and the recommendation efficiency are improved.

Drawings

FIG. 1 is a schematic architecture diagram of a recommendation processing system provided in an embodiment of the present application;

Fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present application;

FIGS. 3A-3C are flow diagrams of a recommendation processing method provided in an embodiment of the present application;

FIG. 4 is an interactive flow diagram of a recommendation processing method according to an embodiment of the present application;

fig. 5 is a schematic diagram of a recommendation flow of a recommendation processing method according to an embodiment of the present application;

fig. 6 is a schematic diagram of a screening flow of the recommendation processing method according to the embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail with reference to the accompanying drawings, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.

In the following description, the terms "first", "second", "third" and the like are merely used to distinguish similar objects and do not represent a specific ordering of the objects, it being understood that the "first", "second", "third" may be interchanged with a specific order or sequence, as permitted, to enable embodiments of the application described herein to be practiced otherwise than as illustrated or described herein.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the present application.

Before further describing embodiments of the present application in detail, the terms and expressions that are referred to in the embodiments of the present application are described, and are suitable for the following explanation.

1) Consumption: the recommended service object provides the service to the object to be serviced with an amount spent, for example, an amount spent by the advertiser in advertising.

2) Entropy weight method: the basic idea of the entropy weight method is to determine objective weights according to the size of index variability. Generally, if the information entropy of a certain index is smaller, the degree of variation of the index is larger, the provided information amount is larger, the function in comprehensive evaluation is also larger, and the weight is also larger. Conversely, the larger the information entropy of a certain index, the smaller the degree of variation of the index value, the smaller the information amount provided, and the smaller the function played in the comprehensive evaluation, and the smaller the weight.

3) Successful advertisement: advertisements with conversion of greater than 30 within 7 days.

4) Conversion amount: the advertiser selects the desired target for advertising, e.g., click, install, activate, register, etc., prior to advertising, and the number of targets achieved is the conversion.

5) Account for successful initiation: there are successful account-based ads, such as the number of successful account-based accounts in the financial industry that the facilitator has ever implemented when recommending advertisers to a customer in the financial industry.

6) The cost is achieved: the cost is the ratio of the actual bid to the target bid, and the cost is between 0.8 and 1.2, representing the cost achievement, wherein the actual bid is the actual consumption of the advertising acquisition conversion, and the target bid is the cost the customer is willing to pay for each conversion before putting the advertisement.

7) And (5) expanding advertisement: and the recharging amount of the advertiser is increased by more than 20% compared with the recharging amount of the previous period in the time period.

In the related art, the selection of the to-be-served object to the recommended service object depends more on the volume of the recommended service object in the current industry and the marketing strength of the recommended service object, and a professional recommendation system is lacked to support the decision of selecting the recommended service object, so that the applicant finds that the manner of selecting the recommended service object in the related art has the following problems when implementing the embodiment of the application:

The first object to be served generally selects the recommended service object only depending on the amount of money of the recommended service object in the industry, so that the index comprehensive capacity of the recommended service object is ignored, and the recommended service object with larger amount of money in the industry may not have excellent index comprehensive capacity, resulting in poor final advertisement delivery effect.

Secondly, the object to be served has many personalized capability requirements for the recommended service object, such as video material making capability, technical service capability, applet construction capability and the like, and propaganda and marketing of the recommended service object can exaggerate the advertising effect and operation capability of the recommended service object, so that the object to be served can hardly distinguish capability differences among different recommended service objects, and the advertising effect can be influenced if the selection is poor.

Third, if the body volume of the object to be served is small, the recommended service object with a large industry share has no effort to serve, so that the selection of the recommended service object is difficult for the small body volume of the object to be served.

In view of the above technical problems, embodiments of the present application provide a recommendation processing method, apparatus, electronic device, computer program product, and computer readable storage medium, which can combine the personalized requirements of index preference of an object to be served with individual scores of objective indexes of a recommended service object to comprehensively consider, and execute recommendation according to the comprehensive considering result.

The recommendation processing method provided by the embodiment of the application can be implemented by various electronic devices, for example, can be implemented by a terminal or a server alone or can be implemented by the terminal and the server cooperatively.

Referring to fig. 1, fig. 1 is a schematic architecture diagram of a recommended processing system provided in an embodiment of the present application, a terminal 400 is connected to a server 200 through a network 300, where the network 300 may be a wide area network or a local area network, or a combination of the two.

In some embodiments, the functions of the recommendation processing system are implemented according to the server 200 and the terminal 400, where the to-be-served object is an advertiser, and the recommended service object is a server, and in the process that the to-be-served object uses the terminal 400, in response to the terminal 400 receiving a trigger operation of the to-be-served object for the recommended function control, the terminal 400 displays an attribute information input interface, in response to the terminal 400 receiving attribute information input by the to-be-served object, the terminal 400 sends the attribute information to the server 200, determines, by the server 200, a comprehensive score of the recommended service object according to the attribute information and individual scores of a plurality of indexes of each recommended service object, performs a first descending order on the plurality of recommended service objects according to the comprehensive score of each recommended service object, the server 200 returns the recommended service object in the first descending order to the terminal 400, and displays the recommended service object on the terminal 400 according to the first descending order.

In other embodiments, when the recommendation processing method provided in the embodiments of the present application is implemented by a terminal alone, the description is given by taking an example that an object to be served is an advertiser, in a process that the object to be served uses the terminal 400, in response to the terminal 400 receiving a trigger operation of the object to be served for a recommendation function control, the terminal 400 displays an attribute information input interface, in response to the terminal 400 receiving attribute information input by the object to be served, determines, by the terminal 400, a composite score of the recommended service object according to the attribute information and individual scores of a plurality of indexes of each recommended service object, performs a first descending order on the plurality of recommended service objects according to the composite score of each recommended service object, and displays the recommended service object on the terminal 400 according to the first descending order ranking result.

In some embodiments, the server 200 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal 400 may be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a smart television, a smart car device, etc., and the terminal 400 may be provided with a client, for example, a video client, a browser client, an information flow client, an image capturing client, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiments of the present application.

In some embodiments, the terminal or the server may implement the recommendation processing method provided in the embodiments of the present application by running a computer program. For example, the computer program may be a native program or a software module in an operating system; may be a local (Native) Application program (APP), i.e., a program that needs to be installed in an operating system to run, such as an advertisement delivery APP; the method can also be an applet, namely a program which can be run only by being downloaded into a browser environment; but also an applet that can be embedded in any APP. In general, the computer programs described above may be any form of application, module or plug-in.

Next, a structure of an electronic device for implementing the recommendation processing method provided in the embodiment of the present application is described, and as before, the electronic device provided in the embodiment of the present application may be the server 200 or the terminal 400 in fig. 1. Referring to fig. 2, fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present application, and the electronic device is taken as a terminal 400 for illustration. The terminal 400 shown in fig. 2 includes: at least one processor 410, a memory 450, at least one network interface 420. The various components in terminal 400 are coupled together by a bus system 440. It is understood that the bus system 440 is used to enable connected communication between these components. The bus system 440 includes a power bus, a control bus, and a status signal bus in addition to the data bus. But for clarity of illustration the various buses are labeled in fig. 2 as bus system 440.

The processor 410 may be an integrated circuit chip having signal processing capabilities such as a general purpose processor, such as a microprocessor or any conventional processor, or the like, a digital signal processor (DSP, digital Signal Processor), or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like.

The user interface 430 includes one or more output devices 431, including one or more speakers and/or one or more visual displays, that enable presentation of the media content. The user interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

Memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard drives, optical drives, and the like. Memory 450 optionally includes one or more storage devices physically remote from processor 410.

Memory 450 includes volatile memory or nonvolatile memory, and may also include both volatile and nonvolatile memory. The non-volatile memory may be read only memory (ROM, read Only Me mory) and the volatile memory may be random access memory (RAM, random Access Memor y). The memory 450 described in the embodiments herein is intended to comprise any suitable type of memory.

In some embodiments, memory 450 is capable of storing data to support various operations, examples of which include programs, modules and data structures, or subsets or supersets thereof, as exemplified below.

An operating system 451 including system programs for handling various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and handling tasks according to hardware;

a network communication module 452 for accessing other electronic devices via one or more (wired or wireless) network interfaces 420, the exemplary network interface 420 comprising: bluetooth, wireless compatibility authentication (WiFi), and universal serial bus (USB, universal Serial Bus), etc.;

a presentation module 453 for enabling presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 (e.g., a display screen, speakers, etc.) associated with the user interface 430;

an input processing module 454 for detecting one or more user inputs or interactions from one of the one or more input devices 432 and translating the detected inputs or interactions.

In some embodiments, the recommendation processing device provided in the embodiments of the present application may be implemented in a software manner, and fig. 2 shows the recommendation processing device 455 stored in the memory 450, which may be software in the form of a program, a plug-in, or the like, including the following software modules: the acquisition module 4551, the index module 4552 and the ranking module 4553 are logical, so that any combination or further splitting may be performed according to the functions implemented, and the functions of the respective modules will be described below.

The recommendation processing method provided by the embodiment of the application will be described with reference to an exemplary application and implementation of the terminal provided by the embodiment of the application.

Referring to fig. 3A, fig. 3A is a schematic flow chart of a service monitoring method according to artificial intelligence according to an embodiment of the present application, and will be described with reference to steps 101-105 shown in fig. 3A.

In step 101, first attribute information of an object to be served is acquired.

As an example, the first attribute information is used to characterize a preference degree of the object to be served for each index of the recommended service object.

As an example, the object to be served is an object that receives a service, for example, with respect to an advertiser and an advertiser, the advertiser belongs to the object to be served, the advertiser belongs to the recommended service object, the first attribute information of the object to be served is obtained in response to the first attribute information of the object to be served input through a man-machine interaction interface in a question-answer form, the first attribute information may also be obtained through historical data of the object to be served, the first attribute information includes a preference degree of the object to be served for each index of the recommended service object, for example, there are A, B, C indexes for measuring the recommended service object, but the object to be served is more important than index a, for example, when a selection operation of the object to be served for index a is received, index a is regarded as a preference index, for example, when a scoring operation of the object to be served for three indexes is received, index a score is 50 score, index B score is 70 score, and index C score is 90 score is regarded as a preference index of the object to be served.

In step 102, individual scores for a plurality of metrics for each recommended service object are determined.

In some embodiments, referring to fig. 3B, fig. 3B is a schematic flow chart of a business monitoring method according to artificial intelligence provided in an embodiment of the present application, and determining individual scores of multiple indicators of each recommended service object in step 102 may be implemented through steps 1021-1023 shown in fig. 3B.

In step 1021, an index value of each index of each recommended service object is determined.

Illustratively, the recommendation service object is an advertising service agent, and the metrics include at least one of: the method comprises the steps of obtaining industry information of an object to be served, and executing at least one of the following steps: determining the consumption number of the starting advertisement of the new advertisement corresponding to the industry information of the recommended service object and the first total consumption number of the new advertisement corresponding to the industry information of the recommended service object, and determining a starting loss index value which is positively correlated with the consumption number of the starting advertisement and negatively correlated with the first total consumption number; determining the number of successful account numbers of the industry information corresponding to the recommended service object and the total number of new account numbers of the industry information corresponding to the recommended service object, and determining a starting speed index value positively correlated with the number of successful account numbers of the starting and negatively correlated with the total number of the new account numbers; determining conversion quantity of industry information corresponding to the recommended service object and exposure quantity of the industry information corresponding to the recommended service object, and determining conversion efficiency index values positively correlated with the conversion quantity and negatively correlated with the exposure quantity; determining a cost achievement advertising consumption number of all advertisements of the industry information corresponding to the recommended service object and a second total consumption number of all advertisements of the industry information corresponding to the recommended service object, and determining a cost achievement rate index value which is positively correlated with the cost achievement advertising consumption number and negatively correlated with the second total consumption number; the consumption number of the expansion advertisements corresponding to the industry information of the recommended service object and the recharging number of the expansion advertisements corresponding to the industry information of the recommended service object are determined, and the expansion capacity index value positively correlated with the consumption number and negatively correlated with the recharging number is determined.

In step 1022, the index value of each index of each recommended service object is normalized, and a normalized index value is obtained.

As an example, first, an index value of each index of each recommended service object is normalized by the formula (1):

wherein Y is _ij A normalized index value X, which is the j index of the i-th recommended service object _ij An index value of the j-th index of the i-th recommended service object, min (X _i ) Is the minimum value, min (X _i ) Is the maximum value of index values of a plurality of indexes of the i-th recommended service object.

In step 1023, for each index, summing the normalized index values of the indexes corresponding to the plurality of recommended service objects to obtain a first summation result, and determining a proportion of the normalized index value of the index corresponding to each recommended service object to occupy the first summation result as an individual score of the index corresponding to each recommended service object.

As an example, the specific gravity of the index value of the j-th index of the i-th recommended service object to the first sum result of the index values of all recommended service objects for the j-th index is calculated by the formula (2):

Wherein P is _ij Is the individual score of the ith recommended service object corresponding to the jth index, Y _ij Is the normalized index value of the j-th index of the i-th recommended service object, and n is the number of recommended service objects.

In step 103, the information entropy of each index is determined according to the individual scores of the plurality of indexes of the plurality of recommended service objects, and the weight of each index is determined according to the information entropy of each index and the first attribute information.

In some embodiments, in step 103, the determining the information entropy of each index according to the individual scores of the multiple indexes of the multiple recommended service objects may be implemented by the following technical solutions: for each index, the following processing is performed: carrying out logarithmic processing on the individual scores of the indexes corresponding to each recommended service object, and multiplying the logarithmic processing result with the individual scores of the indexes corresponding to the recommended service objects; summing the multiplied results of the plurality of recommended service objects to obtain a second summed result; an information entropy of an index that is inversely related to the second summation result is determined.

As an example, the information entropy of each index is calculated by the formula (3):

wherein E is _j Information entropy of j index, P _ij Is the individual score of the ith recommended service object corresponding to the jth index lnP _ij Is the log-processed result of the individual scores of the ith recommended service object corresponding to the jth index,

is the second summation result and n is the number of recommended service objects.

In some embodiments, in step 103, the determining the weight of each index according to the information entropy of each index and the first attribute information may be implemented by the following technical solutions: according to the preference degree of each index, sorting the indexes in a descending order; determining at least one index ranked first as a preference index, and determining the rest indexes as non-preference indexes; the following processing is performed for each preference index: determining the information entropy of the preference index and the optimal information entropy of the corresponding preference index which is positively correlated with the first minimum information entropy and negatively correlated with the second minimum information entropy; the first minimum information entropy is the minimum information entropy in the information entropies of the plurality of indexes, and the second minimum information entropy is the minimum information entropy in the information entropies of the at least one preference index; the following processing is performed for each non-preference index: determining the information entropy of the non-preference index as the optimized information entropy corresponding to the non-preference index; summing the optimized information entropy of the multiple indexes to obtain an information entropy sum; the following processing is performed for each index: and determining the weight of the index which is inversely related to the optimized information entropy of the index and positively related to the information entropy sum.

As an example, when there are 5 indexes, taking the 1 st and 2 nd indexes as preference indexes and the 3 rd, 4 th and 5 th indexes as non-preference indexes as examples, the information entropy value is optimized according to the business experience, the weights of the 1 st and 2 nd indexes are increased, and even one of the 1 st and 2 nd indexes is the maximum value of the weights of the 5 indexes, see formula (4) and formula (5):

Q _j ＝E _j (j＝3,4,5) (5)；

wherein Q is _j Optimized information entropy being the j-th index, E _j Is the information entropy of the 1 st index and the information entropy of the 2 nd index, min (E _j ) Is the minimum value of the information entropy of the 1 st index and the information entropy of the 2 nd index, E _J Is the information entropy of the 3 rd index, the information entropy of the 4 th index and the information entropy of the 5 th index, min (E _J ) Is the minimum value of the information entropy of the 3 rd index, the information entropy of the 4 th index and the information entropy of the 5 th index.

As an example, the weight of each index is calculated by formula (6):

wherein W is _j Is the weight of the j index, Q _j The optimized information entropy of the j index, k is a constant, and Σq _j Is the information entropy sum of the optimized information entropy of all indexes.

In step 104, for each recommended service object, the individual scores of the multiple indexes of the recommended service object are weighted according to the weight of each index, so as to obtain a comprehensive score of the recommended service object.

As an example, the composite score of the index is calculated by equation (7):

wherein S is _i Is the composite score of the ith recommended service object, P _ij Is the individual score of the ith recommended service object corresponding to the jth index, W _j Is the weight of the j-th index. And finally, obtaining comprehensive scores of indexes of each recommended service object, and arranging the plurality of recommended service objects in a descending order according to the comprehensive scores.

In step 105, the plurality of recommended service objects are sorted in a first descending order according to the composite score of each recommended service object, and a recommendation operation is performed on the objects to be served according to the result of the first descending order sorting.

In some embodiments, referring to fig. 3C, fig. 3C is a schematic flow chart of a service monitoring method according to artificial intelligence provided in an embodiment of the present application, and performing a recommendation operation on an object to be served according to a first descending order of sorting results in step 105 may be implemented through steps 1051-1056 shown in fig. 3C.

In step 1051, industry information and rating information for an object to be serviced is obtained.

In step 1052, a plurality of historical objects to be served corresponding to industry information and rating information are queried.

In step 1053, a number of historical recommended service objects for each historical object to be serviced is determined.

In step 1054, the number of historical recommended service objects of the plurality of historical service objects is averaged to obtain an average number.

In step 1055, the top-ranked, average-numbered recommended service objects in the first descending ranking result are obtained.

In step 1056, a recommendation operation is performed on the objects to be served according to the average number of recommended service objects.

As an example, the first M recommended service objects of the score descending result are truncated, and the to-be-served object is recommended to select N recommended service objects, where M is calculated in the following manner, a historical to-be-served object in the same industry and class as the to-be-served object in the historical data is obtained, an average number of historical recommended service objects used by each historical to-be-served object is obtained, where the average number may be an average number of historical recommended service objects selected once, and the average number may also be an average number of historical recommended service objects selected in a period of time.

In some embodiments, when a recommendation operation is performed on the to-be-served object according to the average number of recommended service objects, the to-be-served object may be prompted to select N recommended service objects from M recommended service objects, where N is the number of historical recommended service objects capable of covering 80% of service specific gravity in the historical recommended service objects used by the historical to-be-served object having the same industry and grade as the to-be-served object in the historical data, because in the historical recommended service objects selected by the historical to-be-served object, one part of the historical recommended service objects bears the service specific gravity of the to-be-served object with a large specific gravity, and the other part of the historical recommended service objects can only rob to a small share, so that it is recommended that the number of historical recommended service objects to be covered by 80% of the historical recommended service objects in the historical data is selected by the to-be-served object, and if the number of service merchants is too large, the communication cost of the client is too large.

In some embodiments, the recommended service objects are obtained by screening from candidate recommended service objects, second attribute information of the objects to be served is obtained before individual scores of a plurality of indexes of each recommended service object are determined, and at least one first screening process is performed on the plurality of candidate recommended service objects according to the second attribute information, so that a first screening result is obtained; acquiring third attribute information of an object to be served, and inquiring candidate recommended service objects matched with the third attribute information of the object to be served from the first screening result to form a target screening result set, wherein the third attribute information is used for representing type information of the object to be served; and determining at least one recommended service object matched with the object to be served according to the target screening result set.

As an example, the recommended service objects are obtained by screening from candidate recommended service objects, and when screening is performed, the screening may be performed sequentially as shown in fig. 5, for example, firstly, the screening is performed by an industry preference screening device, then the screening is performed by a client body quantity screening device, then the next screening is determined by the type information, if the next screening is required, the screening is performed by a traffic preference screening device and a support capability screening device, and the final screening result is taken as the recommended service object, so that the schemes from step 101 to step 105 are continuously performed. It is also possible to directly take all candidate recommended service objects as recommended service objects without filtering, so as to continue to execute the schemes of steps 101 to 105.

For example, for the technical solutions of steps 101 to 105, when the index weighted sorting condition is satisfied, the solutions of steps 101 to 105 may be determined, for example, the index weighted sorting condition is that the number of candidate recommended service objects exceeds the number threshold, and the to-be-served object needs to perform multi-index comprehensive consideration on the recommended service objects, for example, the to-be-served object hooks a control of multi-index composite sorting.

In some embodiments, when the second attribute information includes industry information of the object to be served, the first filtering process may be implemented by the following technical scheme: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining total incomes of the candidate recommended service objects, incomes of the candidate recommended service objects corresponding to the industry information and incomes of the corresponding industry information; determining a first industry index which is inversely related to the total income of the candidate recommended service object and is inversely related to the income of the industry information corresponding to the candidate recommended service object; determining a second industry index which is inversely related to the income of the corresponding industry information and is inversely related to the income of the industry information corresponding to the candidate recommended service object; and when the first industry index is larger than the first threshold value and the second industry index is larger than the second threshold value, reserving the candidate recommended service objects.

As an example, the industry preference filter mainly serves to evaluate the deep ploughing degree of the candidate recommended service object in the industry, the first industry index is the ratio of the income of the candidate recommended service object in the industry where the customer is located to the total income of the candidate recommended service object, the second industry index is the income of the candidate recommended service object in the industry where the customer is located to the income of the current industry, when the values of the two indexes are greater than the respective median, the candidate recommended service object enters the next calculation, otherwise, is directly removed from the recommended range.

In some embodiments, when the second attribute information includes at least one level information corresponding to the object to be served, the first filtering process may be implemented by the following technical scheme: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining, for each of the rank information, revenue corresponding to the industry information and rank information for the candidate recommended service object, revenue corresponding to the industry information and rank information, revenue corresponding to the industry information, and revenue corresponding to the industry information for the candidate recommended service object; for each grade information, determining industry information corresponding to the candidate recommended service object and grade information, and grade indexes of positive correlation with the income of the corresponding industry information, negative correlation with the income of the industry information corresponding to the candidate recommended service object and negative correlation with the income of the corresponding industry information and grade information; and when the grade index obtained for any one grade information meets the grade preference condition, carrying out reservation processing on the candidate recommended service object.

As an example, the object-level preference filter is mainly used to evaluate whether the candidate recommended service object can bear the object-level preference, referring to fig. 6, fig. 6 is a schematic diagram of a filtering flow of the recommendation processing method provided in the embodiment of the present application, in which more than 5 million are S-level objects to be served, in which 5 to 1 million are a-level objects to be served, in which 1 to 1 million are B-level objects to be served, in which 1 million to 1 million are C-level objects to be served, in which less than 1 million are D-level objects to be served, the class index of each candidate recommended service object is calculated according to the budget of the object to be served to obtain the class of the object to be served, the numerator of the grade index is the ratio of the income of the candidate recommended service object in the industry where the object to be served is located and the grade where the object to be served is located to the income of the candidate recommended service object in the industry where the object to be served is located, the denominator of the grade index is the ratio of the income of the industry where the object to be served is located and the grade where the object to be served is located to the income of the industry where the object to be served is located, the grades where the object to be served is located, the grade where the object to be served is reduced by one grade, the grade where the object to be served is located is increased by one grade, any result is larger than the corresponding median, the candidate recommended service object enters the next calculation, otherwise, the candidate recommended service object is directly removed from the recommended range.

For example, if the object to be served is a class a, three results are obtained by substituting the class S, the class a and the class B, if the class of the object to be served is a class S, two results of the class S and the class a are required to be calculated, and two results of the class C and the class D are required to be calculated, so that the advantage of calculating adjacent three classes is that not only can the recommended candidate recommended service object be guaranteed to accept the class of the object to be served, but also enough candidate recommended service objects can be guaranteed to be selected by the object to be served.

As an example, at least one first screening process is sequentially performed, and each first screening object is a screening result of a last first screening process, and if the first screening process is a first screening process, the first screening object of the first screening process is an original candidate recommended service object.

In some embodiments, when the type information characterizes that the matched candidate recommended service objects are to be ranked in a non-independent index ranking manner, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining at least one recommended service object matched with the object to be served according to the target screening result set may be achieved by the following technical scheme: determining candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining that a direct recommendation condition is met; and when the direct recommendation condition is met, recommending the service object to be served according to the candidate recommended service object in the target screening result set.

In some embodiments, when the type information characterizes that the matched candidate recommended service objects are to be ranked in an independent index ranking manner, determining at least one recommended service object matched with the object to be served according to the target screening result set may be achieved by the following technical scheme: determining candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization sorts the matched candidate recommended service objects by adopting an independent index sorting mode, determining that an independent index sorting condition is met; when the independent index sorting condition is met, determining independent scores of indexes of corresponding type information of the plurality of recommended service objects, sorting the plurality of recommended service objects according to a second descending order of the independent scores, and executing recommendation operation to the objects to be served according to a second descending order sorting result.

As an example, after at least one first filtering process, query is performed through type information, when the type information characterizes that a non-independent index ordering mode is to be adopted to order matched candidate recommended service objects, and the number of matched candidate recommended service objects does not exceed a number threshold, for example, referring to fig. 5, if the type of the to-be-served object is a pad type, the main appeal of the to-be-served object is not each operation index, the candidate recommended service objects with strong fund capability are directly used as recommended service objects, and are ordered according to a second descending order of fund capability according to the recommended service objects, and then recommendation is performed according to a second descending order ordering result, if the type of the to-be-served object is an operation type, and the to-be-served object is an area to-be-served object, and because the number of the candidate recommended service objects in the area is limited, the candidate recommended service objects in the area are directly used as recommended service objects, and the recommended operations corresponding to the recommended service objects are executed to the to-be-served object.

In some embodiments, when the type information characterizes that the matched candidate recommended service objects are to be ranked in a non-independent index ranking manner and the number of the matched candidate recommended service objects exceeds a number threshold, determining at least one recommended service object matched with the object to be served according to the target screening result set may be achieved by the following technical scheme: and obtaining fourth attribute information of the object to be served, and performing at least one second screening treatment on the target screening result set according to the fourth attribute information to obtain at least one recommended service object matched with the object to be served.

In some embodiments, the fourth attribute information includes at least one flow channel information of the object to be served, and the second filtering process may be implemented by the following technical scheme: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: determining, for each of the traffic channel information, the incomes of the industry information and the traffic channel information corresponding to the candidate recommended service object, the incomes of the industry information corresponding to the candidate recommended service object, and the incomes of the industry information corresponding to the candidate recommended service object; for each flow channel information, determining a flow index which is positively correlated with the incomes of the industry information and the flow channel information corresponding to the candidate recommended service object, is positively correlated with the incomes of the corresponding industry information, is negatively correlated with the incomes of the industry information corresponding to the candidate recommended service object, and is negatively correlated with the incomes of the corresponding industry information and the flow channel information; and when the flow index obtained for each flow channel information meets the channel preference condition, reserving the candidate recommended service object.

As an example, in the traffic preference filter, if the client selects the preferred traffic (optional), the traffic preference filter is triggered to filter, and the advertisement traffic is classified into 4 types: the social dynamic flow, the public platform flow, the video news flow, the alliance flow and the like are respectively calculated according to 4 types of flows, and the social dynamic flow is taken as an example for explanation, wherein the numerator of the flow index is the ratio of the income of the service provider in the industry where the client is located and the income of the service provider in the industry where the client is located, and the denominator of the flow index is the ratio of the income of the industry where the client is located and the income of the industry where the client is located. Depending on the customer's selected preferred traffic, the facilitator will enter the next calculation if there is one greater than the median in the traffic type, otherwise it will be removed directly from the recommended range.

In some embodiments, the second attribute information includes at least one capability requirement information of the object to be served, and the optional second filtering process may be implemented by the following technical solutions: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: and when the candidate recommended service object accords with the capability requirement information, carrying out reservation processing on the candidate recommended service object.

As an example, in the support capability filter, to meet the customer's release requirements, the customer may be provided with a function of performing multiple selections, specifically three capability filters, including: technical facilitator capabilities, i.e. whether the facilitator can provide technical support for the market application program interface, applet creation capabilities, i.e. whether the facilitator can create an applet, video material creation capabilities, i.e. whether the facilitator provides video material creation services, if the client selects an applet creation capability, the facilitator without applet creation capability is removed directly from the recommended range. Otherwise, entering the subsequent judging process. And the screening module finishes the process, and a list of the enclosing service providers is obtained at the moment, namely the recommended service objects are input into the sorting module.

As an example, at least one second screening process is sequentially performed, and each second screening object is a screening result of the last second screening process, and if the second screening process is the first second screening process, the second screening object of the second screening process is an original candidate recommended service object.

According to the embodiment of the invention, the service providers can be screened according to the industry and the volume of the advertisement owner, the flow preference, the operation capability preference and other personalized requirements of the advertisement owner and the objective operation capability and fund capability of the service providers are comprehensively considered, and finally the alternative service providers are recommended to the advertisement owner according to the volume of the advertisement owner, so that the service providers recommended to the advertisement owner can be well matched with the actual conditions and requirements of the advertisement owner in the industry, the volume and various capability items, and the advertisement management method and the advertisement management system can help the advertisement owner to find the best matched service providers, and promote the advertisement effect of the advertisement owner.

In the following, an exemplary application of the embodiments of the present application in a practical application scenario will be described.

In some embodiments, the recommendation processing system is implemented according to a server and a terminal, and the description is given by taking an advertiser as an object to be served and taking a service provider as an example of the object to be recommended, in a process that the object to be served uses the terminal, in response to the terminal receiving a trigger operation of the object to be served for a recommendation function control, the terminal displays an attribute information input interface, in response to the terminal receiving attribute information input by the object to be served, the terminal sends the attribute information to the server, determines comprehensive scores of the recommended service objects according to the attribute information and individual scores of a plurality of indexes of each recommended service object by the server, sorts the recommended service objects in a first descending order according to the comprehensive scores of each recommended service object, returns the recommended service objects in the first descending order sorting result to the terminal, and displays the recommended service objects on the terminal according to the first descending order sorting result.

In some embodiments, referring to fig. 4, fig. 4 is an interactive flow chart of a recommendation processing method provided in the embodiments of the present application, when a client is an advertiser, and after the client creates an advertisement delivery account, a server recommendation function is prompted on an advertisement delivery system, when the client clicks into the server recommendation system, four answer questions need to be answered, three answer questions are selected, the server recommendation system automatically calculates a server most matched with the client according to information (attribute information) input by the client, displays M server lists, and prompts the client with a document to select N servers from among which the client is responsible for advertisement delivery services of the client itself, where M and N are values calculated according to the volume amounts of the client, so that the subsequent client can conduct consultation and engagement through the displayed information of the server.

In some embodiments, referring to fig. 5, the facilitator recommendation system is divided into three modules as a whole: the screening module screens a service provider list which is most in line with the current situation and preference of the client according to the answers (attribute information) of the four necessary questions and the two selected questions input by the client, and the data time range participating in calculation can be selected in the last year.

In some embodiments, the screening module includes an industry preference screener, which is primarily operative to evaluate the deep ploughing of the business by the facilitator, the first industry index being the ratio of the current business of the facilitator to the consumption of the facilitator, the first industry index being the ratio of the income of the customer to the total income of the facilitator, the second industry index being the ratio of the current business of the facilitator to the total consumption of the business, the second industry index being the income of the customer to the income of the current business, the facilitator entering the next calculation when both index values are greater than the respective median, otherwise being directly removed from the recommended range.

In some embodiments, the screening module may further include a customer level preference screener, referring to fig. 6, for evaluating whether the service provider can carry the customer volume, for inputting greater than 5 hundred million as class S customers, for inputting 5 hundred million to 1 hundred million as class a customers, for inputting 1 hundred million to 1 million as class B customers, for inputting 1 million to 1 million as class C customers, for inputting less than 1 million as class D customers, obtaining a customer level according to a customer budget, calculating a level indicator of each service provider, a numerator of the level indicator being a ratio of revenue of the service provider in the industry of the customer and the industry of the customer to revenue of the service provider in the industry of the customer, a denominator of the level indicator being a ratio of revenue of the industry of the customer and the industry of the customer, calculating a level of the customer, and any result being greater than a corresponding median, the service provider entering a next calculation, and otherwise removing directly from the recommended range.

For example, if the customer is a grade A, three results are needed to be obtained by substituting and calculating the grade S, the grade A and the grade B, if the customer grade is the grade S, two results of the grade S and the grade A are needed to be calculated, and two results of the grade C and the grade D are needed to be calculated, so that the advantage of calculating the adjacent three grades is that not only can the recommended service provider be ensured to accept the customer grade, but also enough service providers can be ensured to be selected by the customer.

In some embodiments, if the customer type is a pad type, the customer's primary appeal is not to operate, and the customer is a regional customer, and if the customer is a regional customer, the customer is a signature service provider of the region.

In some embodiments, the screening module may further include a traffic preference filter that is triggered to screen if the customer selects a preferred traffic (optional), the advertisement traffic being classified into 4 categories: the flow indexes are respectively calculated according to the types of flows in 4, such as social dynamic flow, public platform flow, video news flow, alliance flow and the like, and the social dynamic flow is taken as an example for explanation, the numerator of the flow index is the ratio of the income of a service provider in the industry where the client is located and the income of the service provider in the industry where the client is located, and the denominator of the flow index is the ratio of the income of the industry where the client is located and the income of the industry where the client is located. Depending on the customer's selected preferred traffic, the facilitator will enter the next calculation if there is one greater than the median in the traffic type, otherwise it will be removed directly from the recommended range.

In some embodiments, the screening module may further include a support capability screening device, to meet a customer's release requirement, and default no requirement, may provide a function of multiple selection by the customer, specifically providing three capability screening devices, including: technical facilitator capabilities, i.e. whether the facilitator can provide technical support for the market application program interface, applet creation capabilities, i.e. whether the facilitator can create an applet, video material creation capabilities, i.e. whether the facilitator provides video material creation services, if the client selects an applet creation capability, the facilitator without applet creation capability is removed directly from the recommended range. Otherwise, entering the subsequent judging process.

And the screening module finishes the process, and a list of the enclosing service providers is obtained at the moment, namely the recommended service objects are input into the sorting module.

In some embodiments, in the ranking module, the management capabilities of the facilitators are comprehensively ranked according to the index preference of the clients, and comprehensive operation score ranks of the enclosing facilitators are obtained. The operation capability items (indexes) are: the starting amount loss (corresponding to the index value of the starting amount loss), the starting amount speed (corresponding to the index value of the starting amount speed), the conversion efficiency (corresponding to the index value of the conversion efficiency), the cost achievement rate (corresponding to the index value of the cost achievement rate), and the capacity of the extender (corresponding to the index value of the capacity of the extender).

In some embodiments, the bid amount is used to evaluate whether the facilitator is wasting more money from the customer when the advertisement is bid, and the bid amount score is the ratio of the bid advertisement consumption by the facilitator when the customer is in the business to which the new advertisement is to be placed to the total consumption by the facilitator when the customer is in the business to which the new advertisement is to be placed.

In some embodiments, the billing rate is used to evaluate the billing capability of the facilitator, and the billing rate score is the ratio of the number of successful billing accounts of the facilitator in the industry of the customer to the total number of new advertising accounts of the facilitator in the industry of the customer.

In some embodiments, conversion efficiency is used to evaluate the conversion capability of a facilitator, and the conversion efficiency score is the ratio of the conversion by the facilitator in the customer's business to the exposure by the facilitator in the customer's business.

In some embodiments of the present invention, in some embodiments, the cost achievement rate score is the cost of all advertisements of the business where the service provider is located the total consumption of advertisements and all advertisements of the business of the service provider in the client is achieved.

In some embodiments, the capacity score is a consumption amount of the expanded advertisement by the facilitator in the business of the customer and a top-up amount of the expanded advertisement by the facilitator in the business of the customer.

Since the service provider recommendation system is mainly used for new clients, the starting loss and the starting speed are the most important indexes for the new clients, and if only the entropy weight method is used, the weights of the three latter indexes can be far higher than the starting speed and the starting loss, so that the client preference needs to be considered, and the rationality of the weights needs to be ensured.

In some embodiments, the data is the name of the in-range service provider and five columns of operation indexes, and each column of data is normalized by formula (8):

In some embodiments, calculating the index value of the j index of the i-th recommended service object to the specific gravity of the index by formula (9):

wherein P is _ij Is the individual score of the ith recommended service object corresponding to the jth index, Y _ij The j index that is the i-th recommended service objectIs a normalized index value of (1).

In some embodiments, the information entropy of each index is calculated by equation (10):

In some embodiments, optimizing the information entropy value according to the business experience increases the weight of the onset speed and the onset loss, and ensures that one of the weights of the onset speed and the onset loss is a maximum of 5 indexes, see formula (11) and formula (12):

Q _j ＝E _j (j＝3,4,5) (12)；

In some embodiments, the weights are calculated by equation (13):

In some embodiments, the composite score for the indicator is calculated by equation (14):

wherein S is _i Is the composite score of the ith recommended service object, P _ij Is the individual score of the ith recommended service object corresponding to the jth index, W _j Is the weight of the j-th index.

And finally, obtaining comprehensive scores of the operation indexes of each service provider, and arranging the plurality of service providers in a descending order according to the comprehensive scores.

In some embodiments, in the ranking cut-off module, the module cuts off TOP M from the score descending result, and suggests that the client select N servers for delivery, and the calculation method of M is as follows: the number of service providers used by the clients in the industry and the grade of the target clients in the historical data is average, and the calculation method of N is as follows: in the historical data, the number of 80% of consumed service providers can be covered in the industry and the level of the target client, and because one part of service providers bear the consumption of great specific weight of the client and the other part of service providers can only rob to a small share in the service providers, the service providers are recommended to cover 80% of the service providers in the historical data. If the number of service providers is too large, communication costs for clients are excessive.

The recommendation list of the service providers is generated, and the information such as the names of the service providers, the scores and the comprehensive scores of all the capacity items of the service providers, the cooperation mode and the like are finally displayed to the clients.

The embodiment of the application provides a method for recommending service providers for clients with different demands, which defines a service business name list range suitable for the client according to the industry of the client, annual budget, flow preference and supporting capacity, then sequentially sorts target service objects directly according to fund capacity, sorts service providers in a list according to operation capacity of each dimension of the service providers and weights (related to client preference) of the operation capacity of each dimension, finally measures a plurality of service providers ranked at the front according to a client body to recommend the service providers to the client, and recommends the client to select a plurality of service providers in the service providers to perform advertisement delivery.

The method and the device can comprehensively consider the personalized requirements of the index preference of the object to be served and the individual scores of the objective indexes of the recommended service object, so that the recommended service object recommended to the object to be served can be guaranteed to be well matched with the actual situation and the requirements of the object to be served, and the recommendation accuracy and the recommendation efficiency are improved.

It will be appreciated that in the embodiments of the present application, related data such as user information is referred to, and when the embodiments of the present application are applied to specific products or technologies, user permissions or consents need to be obtained, and the collection, use and processing of related data need to comply with related laws and regulations and standards of related countries and regions.

Continuing with the description below of an exemplary architecture implemented as a software module for the recommendation processing device 455 provided in embodiments of the present application, in some embodiments, as shown in fig. 2, the software module stored in the recommendation processing device 455 of the memory 450 may include: the acquiring module 4551 is configured to acquire first attribute information of an object to be served, where the first attribute information is used to characterize a preference degree of the object to be served for each index of a recommended service object; an metrics module 4552 for determining individual scores for a plurality of metrics for each recommended service object; the index module 4552 is further configured to determine an information entropy of each index according to the individual scores of the plurality of indexes of the plurality of recommended service objects, and determine a weight of each index according to the information entropy of each index and the first attribute information; the index module 4552 is further configured to weight, for each recommended service object, the individual scores of the multiple indexes of the recommended service object according to the weight of each index, to obtain a composite score of the recommended service object; the sorting module 4553 is configured to sort the plurality of recommended service objects in a first descending order according to the composite score of each recommended service object, and execute a recommendation operation on the object to be served according to the result of the first descending order sorting.

In some embodiments, the obtaining module 4551 is further configured to obtain second attribute information of the object to be served before determining the individual scores of the multiple indexes of each recommended service object, and perform at least one first screening process on the multiple candidate recommended service objects according to the second attribute information to obtain a first screening result; acquiring third attribute information of an object to be served, and inquiring candidate recommended service objects matched with the third attribute information of the object to be served from the first screening result to form a target screening result set, wherein the third attribute information is used for representing type information of the object to be served; and determining at least one recommended service object matched with the object to be served according to the target screening result set.

In some embodiments, when the second attribute information includes industry information of the object to be served, the acquiring module 4551 is further configured to: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining total incomes of the candidate recommended service objects, incomes of the candidate recommended service objects corresponding to the industry information and incomes of the corresponding industry information; determining a first industry index which is inversely related to the total income of the candidate recommended service object and is inversely related to the income of the industry information corresponding to the candidate recommended service object; determining a second industry index which is inversely related to the income of the corresponding industry information and is inversely related to the income of the industry information corresponding to the candidate recommended service object; and when the first industry index is larger than the first threshold value and the second industry index is larger than the second threshold value, reserving the candidate recommended service objects.

In some embodiments, when the second attribute information includes at least one level information corresponding to the object to be served, the acquiring module 4551 is further configured to: acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object: determining, for each of the rank information, revenue corresponding to the industry information and rank information for the candidate recommended service object, revenue corresponding to the industry information and rank information, revenue corresponding to the industry information, and revenue corresponding to the industry information for the candidate recommended service object; for each grade information, determining industry information corresponding to the candidate recommended service object and grade information, and grade indexes of positive correlation with the income of the corresponding industry information, negative correlation with the income of the industry information corresponding to the candidate recommended service object and negative correlation with the income of the corresponding industry information and grade information; and when the grade index obtained for any one grade information meets the grade preference condition, carrying out reservation processing on the candidate recommended service object.

In some embodiments, the acquiring module 4551 is further configured to: when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining the candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining that a direct recommendation condition is met; and when the direct recommendation condition is met, recommending the service object to be served according to the candidate recommended service object in the target screening result set.

In some embodiments, the acquiring module 4551 is further configured to: when the type information characterization is to sort the matched candidate recommended service objects in an independent index sorting mode, determining the candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served; when the type information characterization sorts the matched candidate recommended service objects by adopting an independent index sorting mode, determining that an independent index sorting condition is met; when the independent index sorting condition is met, determining independent scores of indexes of corresponding type information of the plurality of recommended service objects, sorting the plurality of recommended service objects according to a second descending order of the independent scores, and executing recommendation operation to the objects to be served according to a second descending order sorting result.

In some embodiments, when the type information characterizes that the matching candidate recommended service objects are to be ranked in a non-independent index ranking manner, and the number of matching candidate recommended service objects exceeds a number threshold, the obtaining module 4551 is further configured to: and obtaining fourth attribute information of the object to be served, and performing at least one second screening treatment on the target screening result set according to the fourth attribute information to obtain at least one recommended service object matched with the object to be served.

In some embodiments, the fourth attribute information includes at least one traffic channel information of the object to be served, and the obtaining module 4551 is further configured to: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: determining, for each of the traffic channel information, the incomes of the industry information and the traffic channel information corresponding to the candidate recommended service object, the incomes of the industry information corresponding to the candidate recommended service object, and the incomes of the industry information corresponding to the candidate recommended service object; for each flow channel information, determining a flow index which is positively correlated with the incomes of the industry information and the flow channel information corresponding to the candidate recommended service object, is positively correlated with the incomes of the corresponding industry information, is negatively correlated with the incomes of the industry information corresponding to the candidate recommended service object, and is negatively correlated with the incomes of the corresponding industry information and the flow channel information; and when the flow index obtained for each flow channel information meets the channel preference condition, reserving the candidate recommended service object.

In some embodiments, the acquiring module 4551 is further configured to: acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object: and when the candidate recommended service object accords with the capability requirement information, carrying out reservation processing on the candidate recommended service object.

In some embodiments, the metrics module 4552 is further to: determining an index value of each index of each recommended service object; performing standardization processing on the index value of each index of each recommended service object to obtain a standardized index value; and summing the standardized index values of the indexes corresponding to the plurality of recommended service objects according to each index to obtain a first summation result, and determining the proportion of the standardized index value of the index corresponding to each recommended service object to occupy the first summation result as an independent score of the index corresponding to each recommended service object.

In some embodiments, the metrics module 4552 is further to: for each index, the following processing is performed: carrying out logarithmic processing on the individual scores of the indexes corresponding to each recommended service object, and multiplying the logarithmic processing result with the individual scores of the indexes corresponding to the recommended service objects; summing the multiplied results of the plurality of recommended service objects to obtain a second summed result; an information entropy of an index that is inversely related to the second summation result is determined.

In some embodiments, the metrics module 4552 is further to: according to the preference degree of each index, sorting the indexes in a descending order; determining at least one index ranked first as a preference index, and determining the rest indexes as non-preference indexes; the following processing is performed for each preference index: determining an information entropy positively correlated with the preference index and a first minimum information entropy, and an optimized information entropy negatively correlated with a second minimum information entropy; the first minimum information entropy is the minimum information entropy in the information entropies of the plurality of indexes, and the second minimum information entropy is the minimum information entropy in the information entropies of the at least one preference index; the following processing is performed for each non-preference index: determining the information entropy of the non-preference index as the optimized information entropy corresponding to the non-preference index; summing the optimized information entropy of the multiple indexes to obtain an information entropy sum; the following processing is performed for each index: and determining the weight of the index which is inversely related to the optimized information entropy of the index and positively related to the information entropy sum.

In some embodiments, the ordering module 4553 is further to: acquiring industry information and grade information of an object to be served; inquiring a plurality of historical objects to be served corresponding to industry information and grade information; determining the number of historical recommended service objects of each historical object to be serviced; averaging the number of the historical recommended service objects of the plurality of historical objects to be serviced to obtain an average number; acquiring the recommended service objects which are ordered in the first descending order and meet the average number and are positioned in front in the first descending order; and according to the recommended service objects which meet the average number, performing recommendation operation on the objects to be served.

Embodiments of the present application provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the recommended processing method according to the embodiment of the application.

The present embodiments provide a computer readable storage medium having stored therein executable instructions that, when executed by a processor, cause the processor to perform the recommended processing method provided by the embodiments of the present application, for example, as shown in fig. 3A-3C.

In some embodiments, the computer readable storage medium may be FRAM, ROM, PROM, EP ROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; but may be a variety of devices including one or any combination of the above memories.

In some embodiments, the executable instructions may be in the form of programs, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and they may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

As an example, the executable instructions may, but need not, correspond to files in a file system, may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a hypertext markup language (HTML, hyper Text Markup Language) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).

As an example, executable instructions may be deployed to be executed on one computing device or on multiple computing devices located at one site or, alternatively, distributed across multiple sites and interconnected by a communication network.

In summary, according to the embodiment of the application, the personalized requirements of the index preference of the object to be served and the objective index individual scoring of the recommended service object can be combined for comprehensive consideration, so that the recommended service object recommended to the object to be served can be ensured to well match the actual situation and the requirements of the object to be served, and the recommendation accuracy and the recommendation efficiency are improved.

The foregoing is merely exemplary embodiments of the present application and is not intended to limit the scope of the present application. Any modifications, equivalent substitutions, improvements, etc. that are within the spirit and scope of the present application are intended to be included within the scope of the present application.

Claims

1. A recommendation processing method, the method comprising:

2. The method of claim 1, wherein prior to determining the individual scores for the plurality of metrics for each of the recommended service objects, the method further comprises:

acquiring second attribute information of the object to be served, and performing at least one first screening treatment on a plurality of candidate recommended service objects according to the second attribute information to obtain a first screening result;

Acquiring third attribute information of the object to be served, and inquiring candidate recommended service objects matched with the third attribute information of the object to be served from the first screening result to form a target screening result set, wherein the third attribute information is used for representing type information of the object to be served;

and determining at least one recommended service object matched with the object to be served according to the target screening result set.

3. The method according to claim 2, wherein when the second attribute information includes industry information of the object to be served, the first filtering process includes:

acquiring a first screening object of the first screening process, and executing the following processes for each candidate recommended service object in the first screening object:

determining total incomes of the candidate recommended service objects, incomes of the candidate recommended service objects corresponding to the industry information and incomes of the candidate recommended service objects corresponding to the industry information;

determining a first industry index that is inversely related to the total revenue of the candidate recommended service object and is positively related to the revenue of the industry information corresponding to the candidate recommended service object;

Determining a second industry index that is inversely related to revenue corresponding to the industry information and that is positively related to revenue corresponding to the industry information for the candidate recommended service object;

and when the first industry index is larger than a first threshold value and the second industry index is larger than a second threshold value, reserving the candidate recommended service object.

4. The method according to claim 2, wherein when the second attribute information includes at least one level information corresponding to the object to be served, the first filtering process includes:

determining, for each of the ranking information, the incomes of the candidate recommended service object corresponding to the industry information and the ranking information, the incomes of the candidate recommended service object corresponding to the industry information;

for each of the ranking information, determining a positive correlation of revenue for the industry information and the ranking information corresponding to the candidate recommended service object, a positive correlation of revenue for the industry information, a negative correlation of revenue for the industry information corresponding to the candidate recommended service object, and a ranking index of negative correlation of revenue for the industry information and the ranking information;

And when the grade index obtained for any one grade information meets the grade preference condition, carrying out reservation processing on the candidate recommended service object.

5. The method according to claim 2, wherein the method further comprises:

when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining at least one recommended service object matched with the object to be served according to the target screening result set, including:

determining candidate recommended service objects in the target screening result set as at least one recommended service object matched with the object to be served;

the method further comprises the steps of:

when the type information characterization is to sort the matched candidate recommended service objects in a non-independent index sorting mode, and the number of the matched candidate recommended service objects does not exceed a number threshold, determining that a direct recommendation condition is met;

and when the direct recommendation condition is met, according to the candidate recommended service objects in the target screening result set, executing recommendation operation to the to-be-served objects.

6. The method of claim 2, wherein the step of determining the position of the substrate comprises,

when the type information characterization is to sort the matched candidate recommended service objects in an independent index sorting mode, determining at least one recommended service object matched with the object to be served according to the target screening result set, wherein the method comprises the following steps:

the method further comprises the steps of:

when the type information characterization sorts the matched candidate recommended service objects in an independent index sorting mode, determining that an independent index sorting condition is met;

and when the independent index sorting condition is met, determining independent scores of indexes of the plurality of recommended service objects corresponding to the type information, sorting the plurality of recommended service objects in a second descending order according to the independent scores, and executing recommendation operation to the objects to be served according to a second descending order sorting result.

7. The method of claim 2, wherein the step of determining the position of the substrate comprises,

when the type information characterizes that the matched candidate recommended service objects are to be sequenced in a non-independent index sequencing mode and the number of the matched candidate recommended service objects exceeds a number threshold, determining at least one recommended service object matched with the object to be served according to the target screening result set comprises the following steps:

And obtaining fourth attribute information of the object to be served, and performing at least one second screening treatment on the target screening result set according to the fourth attribute information to obtain at least one recommended service object matched with the object to be served.

8. The method of claim 7, wherein the fourth attribute information includes at least one traffic channel information of the object to be served, the second filtering process comprising, at any one time:

acquiring a second screening object of the second screening process, and executing the following processes for each candidate recommended service object in the second screening object:

determining, for each of the traffic channel information, the incomes of the candidate recommended service object corresponding to the industry information and the traffic channel information, the incomes of the candidate recommended service object corresponding to the industry information;

for each piece of traffic channel information, determining a traffic index positively correlated with the incomes of the industry information and the traffic channel information corresponding to the candidate recommended service object, positively correlated with the incomes of the industry information, negatively correlated with the incomes of the industry information corresponding to the candidate recommended service object, and negatively correlated with the incomes of the industry information and the traffic channel information;

And when the flow index obtained for each piece of flow channel information meets the channel preference condition, carrying out reservation processing on the candidate recommended service object.

9. The method of claim 7, wherein the second attribute information includes at least one capability requirement information of the object to be served, the second filtering process comprising, at any one time:

and when the candidate recommended service object accords with each piece of capability requirement information, carrying out reservation processing on the candidate recommended service object.

10. The method of claim 1, wherein said determining individual scores for a plurality of said metrics for each of said recommended service objects comprises:

determining an index value of each of the indexes of each of the recommended service objects;

performing standardization processing on the index value of each index of each recommended service object to obtain a standardized index value;

and summing the standardized index values of the indexes corresponding to the plurality of recommended service objects according to each index to obtain a first summation result, and determining the proportion of the standardized index value of the index corresponding to each recommended service object to occupy the first summation result as an independent score of the index corresponding to each recommended service object.

11. The method of claim 1, wherein said determining the entropy of each of said metrics based on individual scores of a plurality of said metrics for a plurality of said recommended service objects comprises:

for each of the indices, the following processing is performed:

carrying out logarithmic processing on the individual scores of the indexes corresponding to each recommended service object, and multiplying the logarithmic processing result with the individual scores of the indexes corresponding to the recommended service objects;

carrying out summation processing on the multiplication processing results of the plurality of recommended service objects to obtain a second summation result;

an information entropy of the indicator that is inversely related to the second summation result is determined.

12. The method of claim 1, wherein determining the weight of each index based on the information entropy of each index and the first attribute information comprises:

sorting the indexes in a descending order according to the preference degree of each index;

determining at least one index which is ranked first as a preference index, and determining the rest indexes as non-preference indexes;

the following processing is performed for each of the preference indicators: determining an optimized information entropy positively correlated with the information entropy of the preference index and the first minimum information entropy and negatively correlated with the second minimum information entropy;

Wherein the first minimum information entropy is the minimum information entropy in the information entropy of a plurality of indexes, and the second minimum information entropy is the minimum information entropy in the information entropy of at least one preference index;

the following processing is performed for each of the non-preference indicators: determining the information entropy of the non-preference index as the optimized information entropy corresponding to the non-preference index;

summing the optimized information entropy of the indexes to obtain information entropy sum;

the following processing is performed for each of the indices: a weight of the index that is inversely related to an optimized information entropy of the index and positively related to the information entropy sum is determined.

13. The method of claim 1, wherein the performing a recommendation operation to the object to be served according to the first descending order of ranking results comprises:

acquiring industry information and grade information of the object to be served;

inquiring a plurality of historical objects to be served corresponding to the industry information and the grade information;

determining the number of historical recommended service objects of each historical object to be serviced;

averaging the number of the historical recommended service objects of the historical objects to be serviced to obtain an average number;

Acquiring the recommended service objects which are ranked ahead and meet the average number in the first descending ranking result;

and according to the recommended service objects which meet the average number, executing recommendation operation to the objects to be served.

14. A recommendation processing apparatus, the apparatus comprising:

15. An electronic device, the electronic device comprising:

a memory for storing executable instructions;

a processor for implementing the recommended processing method of any one of claims 1 to 13 when executing executable instructions stored in the memory.

16. A computer-readable storage medium storing executable instructions that when executed by a processor implement the recommended processing method of any of claims 1 to 13.

17. A computer program product comprising a computer program or instructions which, when executed by a processor, implements the recommended processing method of any of claims 1 to 13.