Nothing Special   »   [go: up one dir, main page]

CN111784455A - Article recommendation method and recommendation equipment - Google Patents

Article recommendation method and recommendation equipment Download PDF

Info

Publication number
CN111784455A
CN111784455A CN202010615163.6A CN202010615163A CN111784455A CN 111784455 A CN111784455 A CN 111784455A CN 202010615163 A CN202010615163 A CN 202010615163A CN 111784455 A CN111784455 A CN 111784455A
Authority
CN
China
Prior art keywords
user
recommended
behavior
item
article
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010615163.6A
Other languages
Chinese (zh)
Other versions
CN111784455B (en
Inventor
田植良
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN202010615163.6A priority Critical patent/CN111784455B/en
Publication of CN111784455A publication Critical patent/CN111784455A/en
Application granted granted Critical
Publication of CN111784455B publication Critical patent/CN111784455B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0631Item recommendations
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241Advertisements
    • G06Q30/0251Targeted advertisements
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0641Shopping interfaces
    • G06Q30/0643Graphical representation of items or shoppers

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the application provides an article recommendation method and recommendation equipment, and relates to the technical field of artificial intelligence, wherein the method comprises the following steps: acquiring attribute information and a behavior sequence of a user to be recommended, and then determining preference characteristics of the user to be recommended for the article according to the attribute information and the behavior sequence. And determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item. Since the historical operation time in the behavior sequence comprises the positive conversion behavior event and the negative conversion behavior event, the positive conversion behavior event is positively correlated with the preference characteristic, and the negative conversion behavior event is negatively correlated with the preference characteristic, when the attribute information of the user to be recommended and the behavior sequence are combined to extract the characteristics of the user, more comprehensive user preference characteristics can be obtained, and the accuracy of recommending articles for the user based on the user preference characteristics is improved.

Description

Article recommendation method and recommendation equipment
Technical Field
The embodiment of the application relates to the technical field of artificial intelligence, in particular to an article recommendation method and recommendation equipment.
Background
With the continuous development of online shopping platforms, recommendation systems have become irreplaceable important components in e-commerce. The recommendation system can learn the hidden preference information in the user historical behaviors, so that the shopping behaviors of the user are further predicted, customers are helped to select satisfied commodities, and the income of an e-commerce platform is promoted to be improved.
At present, many recommendation systems perform static modeling on a user according to basic information such as age and gender of the user, and then recommend commodities to the user by using the obtained model. The method does not consider users with the same basic information, and the preferences of the users may be different, so that the users may be recommended with unnecessary commodities, and the recommendation accuracy is low.
Disclosure of Invention
The embodiment of the application provides an article recommendation method and recommendation equipment, which are used for improving the accuracy of article recommendation.
In one aspect, an embodiment of the present application provides an item recommendation method, where the method includes:
acquiring attribute information and a behavior sequence of a user to be recommended, wherein the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each article according to historical operation time, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events;
determining preference characteristics of the user to be recommended for an article according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics;
and determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
In one aspect, an embodiment of the present application provides a recommendation device, where the recommendation device includes:
the system comprises an acquisition module, a recommendation module and a recommendation module, wherein the acquisition module is used for acquiring attribute information and a behavior sequence of a user to be recommended, the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each article according to historical operation time, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events;
the processing module is used for determining preference characteristics of the user to be recommended for the article according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics;
and the recommending module is used for determining the target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
In one aspect, an embodiment of the present application provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the item recommendation method when executing the program.
In one aspect, embodiments of the present application provide a computer-readable storage medium storing a computer program executable by a computer device, wherein the program, when executed on the computer device, causes the computer device to execute steps of an item recommendation method.
In the embodiment of the application, as the historical operation events in the behavior sequence comprise positive conversion behavior events and negative conversion behavior events, the positive conversion behavior events are positively correlated with the preference features, and the negative conversion behavior events are negatively correlated with the preference features, when the attribute information of the user to be recommended and the behavior sequence are combined to perform feature extraction on the user, more comprehensive user preference features can be obtained, and then a user image with higher accuracy is constructed. Furthermore, when the articles are recommended to the user according to the user preference characteristics and the article characteristics of the articles, the article recommendation accuracy can be effectively improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
Fig. 1 is a schematic diagram of a system architecture according to an embodiment of the present application;
fig. 2 is a schematic flowchart of an article recommendation method according to an embodiment of the present application;
fig. 3a is a schematic diagram of a recommendation page provided in an embodiment of the present application;
FIG. 3b is a schematic diagram of a recommendation page provided in an embodiment of the present application;
FIG. 3c is a schematic diagram of a recommendation page provided in an embodiment of the present application;
fig. 4 is a flowchart illustrating a method for extracting a preference feature according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a convolutional neural network according to an embodiment of the present disclosure;
FIG. 6a is a timing diagram of an embodiment of the present application;
FIG. 6b is a timing diagram of an embodiment of the present application;
FIG. 6c is a timing diagram of an embodiment of the present application;
FIG. 7a is a timing diagram of an embodiment of the present application;
FIG. 7b is a timing diagram of an embodiment of the present application;
FIG. 7c is a timing diagram of an embodiment of the present application;
fig. 8 is a schematic flowchart of an article recommendation method according to an embodiment of the present application;
fig. 9 is a schematic structural diagram of a recommendation device according to an embodiment of the present application;
fig. 10 is a schematic structural diagram of a computer device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more clearly apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
For convenience of understanding, terms referred to in the embodiments of the present invention are explained below.
Artificial Intelligence (AI): the method is a theory, method, technology and application system for simulating, extending and expanding human intelligence by using a digital computer or a machine controlled by the digital computer, sensing the environment, acquiring knowledge and obtaining the best result by using the knowledge. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making.
The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
Natural Language Processing (NLP): is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that enable efficient communication between humans and computers using natural language. Natural language processing is a science integrating linguistics, computer science and mathematics. Therefore, the research in this field will involve natural language, i.e. the language that people use everyday, so it is closely related to the research of linguistics. Natural language processing techniques typically include text processing, semantic understanding, machine translation, robotic question and answer, knowledge mapping, and the like. In the embodiment of the application, the user characteristics, the article characteristics and the like are converted into the characteristic vectors and the like by using the NLP technology.
Machine Learning (ML) is a multi-domain cross discipline, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning. For example, in the embodiment of the application, the preference characteristics of the user are extracted by machine learning, and then the item is recommended to the user based on the preference characteristics.
A convolutional neural network: (Convolutional Neural Networks, CNN) is a type of feedforward Neural network that contains convolution calculations and has a deep structure, and is one of the representative algorithms of deep learning (deep learning). Convolutional neural networks have a characteristic learning ability, and can perform translation invariant classification on input information according to a hierarchical structure thereof, and are also called translation invariant artificial neural networks.
A recurrent neural network: (Current Neural networks, RNN) are a type of Recurrent Neural Network that takes sequence data as input, recurses in the evolution direction of the sequence, and all nodes (cyclic units) are connected in a chain manner, wherein Bidirectional Recurrent Neural networks (Bi-RNN) and Long-Short-Term memory networks (LSTM) are common Recurrent Neural networks. The recurrent neural network has memory, parameter sharing and graph completion (training completion), and thus has certain advantages in learning the nonlinear characteristics of a sequence. The recurrent neural network has applications in Natural Language Processing (NLP), such as speech recognition, Language modeling, machine translation, and other fields, and is also used for various time series predictions.
Timing diagram neural network: the timing diagram neural network is added with a process of processing timing information on the basis of a Graph neural network (GCN). The graph neural network is a vector representation of an output graph with a graph structure as an input. In the graph neural network, nodes in an input graph are represented as vectors, and edges of the input graph represent incidence relations among the nodes. Each node in the neural network of the timing diagram can have a time label besides corresponding to a vector.
Article: broadly refers to various types of products such as merchandise, advertising, etc.
The user to be recommended: the user who needs to recommend the item is referred to, the user to be recommended can be a user in the sample user or a new user, and any user who needs to recommend the item can be regarded as the user to be recommended.
The following is a description of the design concept of the embodiments of the present application.
At present, many recommendation systems perform static modeling on a user according to basic information such as age and gender of the user, and then recommend commodities to the user by using the obtained model. This method does not take into account users whose basic information is the same, and their preferences may be different. For example, a 20-year-old woman may like high-heeled shoes or sneakers, and thus when recommending goods to a user based on basic information of the user, goods that are not needed by the user may be recommended, which results in a low recommendation accuracy.
The user's preference may be reflected in the user's operation behavior, for example, when the user purchases a certain product, it indicates that the user likes the product with a high probability. For another example, when a user collects a certain item, it indicates that the user is interested in the item. At this time, the recommendation system may consider recommending the goods purchased or collected by the user or the like to the user. However, the operation behavior of the user is not all that indicates that the user likes a certain product, for example, when the user returns a product after purchasing a certain product, it indicates that the product purchased by the user may not be suitable for the user. At this time, the recommendation system may consider not recommending the goods returned by the user or the like to the user. In view of this, an embodiment of the present application provides an item recommendation method, including: the method comprises the steps of firstly obtaining attribute information and a behavior sequence of a user to be recommended, wherein the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each article according to historical operation time, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events. And then determining the preference characteristics of the user to be recommended for the article according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics. And then determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
In the embodiment of the application, as the historical operation events in the behavior sequence comprise positive conversion behavior events and negative conversion behavior events, the positive conversion behavior events are positively correlated with the preference features, and the negative conversion behavior events are negatively correlated with the preference features, when the attribute information of the user to be recommended and the behavior sequence are combined to perform feature extraction on the user, more comprehensive user preference features can be obtained, and then a user image with higher accuracy is constructed. Furthermore, when the articles are recommended to the user according to the user preference characteristics and the article characteristics of the articles, the article recommendation accuracy can be effectively improved.
After introducing the design concept of the embodiment of the present application, some simple descriptions are provided below for application scenarios to which the technical solution of the embodiment of the present application can be applied, and it should be noted that the application scenarios described below are only used for describing the embodiment of the present application and are not limited. In specific implementation, the technical scheme provided by the embodiment of the application can be flexibly applied according to actual needs.
Scene one, shopping recommendation scene
When recommending commodities to a user to be recommended, firstly acquiring attribute information and a behavior sequence of the user to be recommended, wherein the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each commodity according to historical operation time, the attribute information of the user to be recommended comprises gender, age, location and the like, and the historical operation events comprise purchase, return, order cancellation, collection cancellation, purchase addition and the like. The historical operation events comprise positive conversion action events and negative conversion action events, wherein the positive conversion events comprise purchase, collection, purchase adding and the like, and the negative conversion action events comprise goods return, order cancellation, collection cancellation and the like. And determining the preference characteristics of the user to be recommended for the commodity according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics. And then determining the target commodity recommended to the user to be recommended from each candidate commodity according to the preference characteristic of the user to be recommended for the commodity and the commodity characteristic of each candidate commodity.
Scene two, advertisement recommendation scene
When the advertisements are recommended to the user to be recommended, firstly, attribute information and a behavior sequence of the user to be recommended are obtained, the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each advertisement according to historical operation time, wherein the attribute information of the user to be recommended comprises gender, age, location and the like, and the historical operation events comprise clicking, browsing, consulting, purchasing, returning goods, canceling orders, collecting, canceling collecting, adding and purchasing and the like. The historical operation events comprise positive conversion action events and negative conversion action events, wherein the positive conversion events comprise clicking, browsing, consulting, purchasing, collecting, purchasing and the like, and the negative conversion action events comprise returning, canceling orders, canceling collections and the like. And determining the preference characteristics of the user to be recommended for the advertisement according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics. And then determining the target advertisement recommended to the user to be recommended from each candidate advertisement according to the preference characteristics of the user to be recommended for the advertisement and the advertisement characteristics of each candidate advertisement.
Reference is made to fig. 1, which is a system architecture diagram of an item recommendation method according to an embodiment of the present application. The architecture comprises at least a terminal device 101 and a server 102.
The terminal device 101 may have a client installed therein, where the client may be a web page version client or a client pre-installed in the terminal device 101, and the client in this embodiment may be an e-commerce shopping client, or a financial client, or any type of client capable of delivering an advertisement. The attribute information of the user to be recommended can be obtained from the registration information of the user to be recommended in the client, and the behavior sequence of the user to be recommended can be obtained from the operation behavior data of the user to be recommended in the client. Terminal device 101 may include one or more processors 1011, memory 1012, I/O interface 1013 to interact with the buried point server 103, and display panel 1014, among other things. The terminal device 101 may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, and the like.
The server 102 may be a background server of the client and provide corresponding services for the client, and the server 102 may include one or more processors 1021, a memory 1022, and an I/O interface 1023 for interaction with the terminal device 101. In addition, server 102 may also configure database 1024. The server 102 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like. The terminal device 101 and the server 102 may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein.
The recommendation device may be the terminal device 101 or the server 102.
In the first case, the recommendation device is the terminal device 101.
When a user to be recommended triggers an article recommendation service in a client of the terminal device 101, the terminal device 101 acquires attribute information and a behavior sequence of the user to be recommended from the server 102, and then determines preference characteristics of the user to be recommended for an article according to the attribute information and the behavior sequence of the user to be recommended. And then determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item. The terminal device 101 presents the target item at the client.
In the second case, the recommendation device is the server 102.
When a user to be recommended triggers an article recommendation service in a client of the terminal device 101, the terminal device 101 sends a recommendation request to the server 102, the server 102 acquires attribute information and a behavior sequence of the user to be recommended from a database according to the recommendation request, and then determines preference characteristics of the user to be recommended for an article according to the attribute information and the behavior sequence of the user to be recommended. And then determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item. The server 102 sends the related information of the target item to the terminal device 101, and the terminal device 101 displays the target item at the client.
Based on the system architecture diagram shown in fig. 1, an embodiment of the present application provides a flow of an item recommendation method, as shown in fig. 2, where the flow of the method is executed by a recommendation device, which may be the terminal device 101 or the server 102 shown in fig. 1, and includes the following steps:
step S201, acquiring attribute information and a behavior sequence of a user to be recommended.
Specifically, the attribute information of the user to be recommended may be obtained from registration information of the user to be recommended, and the attribute information of the user to be recommended at least includes two types of information:
the first type is numerical type, namely information described by numbers, such as age, birth year, month and day, account registration time and the like, and the information can be directly used as numerical characteristics of the user to be recommended to perform characteristic extraction on the user to be recommended.
The second category is an option category, namely attributes selected when the user to be recommended registers, such as sex, male or female, and place, Beijing, Shanghai, and the like. All options of each option class information are coded, then 1 is filled in the dimension which accords with the selection of the user to be recommended, 0 is filled in the other dimensions, a vector feature is formed, and therefore the option class feature is converted into a numerical feature, and then feature extraction is carried out on the user to be recommended.
The behavior sequence is obtained by sequencing the historical operation events of the user to be recommended aiming at each article according to the historical operation time, wherein the historical operation events comprise positive conversion behavior events and negative conversion behavior events.
Specifically, in the shopping recommendation scene, the historical operation events include purchase, return, cancel orders, collect, cancel collect, buy, and the like, wherein the positive conversion events are purchase, collect, buy, and the like, and the negative conversion behavior events are return, cancel orders, cancel collect, and the like. In the advertisement recommendation scene, the historical operation events comprise clicking, browsing, consulting, purchasing, returning goods, canceling orders, collecting, canceling collecting, adding purchasing and the like, wherein positive conversion events comprise clicking, browsing, consulting, purchasing, collecting, adding purchasing and the like, and negative conversion action events comprise returning goods, canceling orders, canceling collecting and the like.
Step S202, determining preference characteristics of the user to be recommended for the article according to the attribute information and the behavior sequence of the user to be recommended.
Specifically, positive conversion behavior events are positively correlated with the preference characteristics, and negative conversion behavior events are negatively correlated with the preference characteristics. For example, it is set that a user to be recommended purchases a commodity a and a commodity B on a shopping client, the user to be recommended leaves the commodity a after receiving the commodity a and the commodity B, and returns the commodity B to a merchant, then a positive conversion behavior event of the user to be recommended is to purchase the commodity a and purchase the commodity B, a negative conversion behavior event of the user to be recommended is to return the commodity B to the merchant, and since the positive conversion behavior event is positively correlated with the preference feature and the negative conversion behavior event is negatively correlated with the preference feature, the obtained preference feature of the user to be recommended for the commodity is the preferred commodity a, and the commodity B may not be suitable for the user to be recommended.
In addition, the behavior sequence is obtained by sequencing the historical operation events of the user to be recommended aiming at each article according to the historical operation time, so the preference characteristic of the user to be recommended along with the change of time can be obtained according to the behavior sequence. For example, if the user to be recommended purchases the commodity C one year ago, collects the commodity B one half year ago, and purchases the commodity F one week ago, it indicates that the user to be recommended prefers the commodity B and the commodity C before, and prefers the commodity F currently, so that the user to be recommended needs the commodity F or the commodity of the same type as the commodity F currently.
Step S203, determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
In a specific implementation, each candidate item is all items that can be recommended to the user, such as all goods on sale, all advertisements to be delivered, and the like. And if the recommendation equipment is terminal equipment, the terminal equipment displays the target item at the client after determining the target item recommended to the user to be recommended. If the recommending device is a server, after the server determines the target object recommended to the user to be recommended, the server sends the relevant information of the target object to the terminal device, and the terminal device displays the target object at the client.
Illustratively, as shown in fig. 3a, when it is determined that the target products recommended to the user to be recommended are "pineapple" and "grape", the recommendation device preferentially displays purchase links of "pineapple" and "grape" in a recommendation page of a fruit category of the shopping client, for example, the purchase links of "pineapple" and "grape" are displayed at the uppermost end of the recommendation page, and purchase links of other recommended fruits are displayed at the lower end of the recommendation page, for example, the purchase links of "banana" and "strawberry" are displayed at the lower end of the recommendation page.
Illustratively, as shown in fig. 3b, when it is determined that the target goods recommended to the user to be recommended are "mouse" and "keyboard", the recommendation device preferentially displays the purchase links of the "mouse" and the "keyboard" in the recommendation page of the shopping client appliance category, for example, the purchase links of the "mouse" and the "keyboard" are displayed at the uppermost end of the recommendation page, and the purchase links of other recommended appliances are displayed at the lower end of the recommendation page, for example, the purchase links of "earphone" are displayed at the lower end of the recommendation page.
Illustratively, as shown in fig. 3c, when it is determined that the target advertisement recommended to the user to be recommended is an automobile advertisement, the recommendation device displays the automobile advertisement in a circle of friends of the instant messaging client. The user to be recommended can click on the advertisement picture to view the advertisement or enter a purchase page.
In the embodiment of the application, as the historical operation events in the behavior sequence comprise positive conversion behavior events and negative conversion behavior events, the positive conversion behavior events are positively correlated with the preference features, and the negative conversion behavior events are negatively correlated with the preference features, when the attribute information of the user to be recommended and the behavior sequence are combined to perform feature extraction on the user, more comprehensive user preference features can be obtained, and then a user image with higher accuracy is constructed. Furthermore, when the articles are recommended to the user according to the user preference characteristics and the article characteristics of the articles, the article recommendation accuracy can be effectively improved.
Optionally, in the step S202, determining the preference characteristic of the to-be-recommended user for the item according to the attribute information and the behavior sequence of the to-be-recommended user, specifically including the following steps, as shown in fig. 4:
step S401, extracting features of attribute information of a user to be recommended by adopting a convolutional neural network, and obtaining a first feature vector of the user to be recommended.
Specifically, the attribute information of the user to be recommended comprises account registration time, gender, age, location and the like, and before the attribute information of the user to be recommended is input into the convolutional neural network for feature extraction, the attribute information of the user to be recommended is converted into numerical features so that the convolutional neural network can perform feature extraction on the attribute information of the user to be recommended. In specific implementation, since attribute information such as registration time, age and the like is described by using numbers, the attribute information can be directly used as numerical characteristics of the user to be recommended. According to attribute information of character descriptions such as gender and places, a coding mode is adopted in the embodiment of the application to convert the attribute information into numerical characteristics.
As shown in fig. 5, the convolutional neural network includes at least a convolutional layer, an activation layer, and a pooling layer. Inputting the attribute information of the user to be recommended into a convolutional layer, and performing feature extraction on the attribute information of the user to be recommended by the convolutional layer through convolutional core, wherein the convolutional core in the convolutional layer can be one or more. And inputting the features obtained by convolution into an activation layer, wherein the activation layer adopts an activation function to carry out nonlinear adjustment on the features, and then inputting the features into a pooling layer. And the pooling layer performs down-sampling operation on the adjusted features to obtain a first feature vector of the user to be recommended.
And S402, performing feature extraction on the behavior sequence by adopting a timing sequence neural network to obtain a second feature vector of the user to be recommended.
Specifically, the timing Graph neural Network is based on a Graph neural Network (GCN), and adds a processing procedure for timing information. The graph neural network is a vector representation of an output graph with a graph structure as an input. In the graph neural network, nodes in an input graph are represented as vectors, and edges of the input graph represent incidence relations among the nodes. Each node in the neural network of the timing diagram can have a time label besides corresponding to a vector.
In the embodiment of the application, the timing graph neural network is obtained by training with the positive conversion behavior event of the sample user as a positive sample and the negative conversion behavior event of the sample user as a negative sample. In specific implementation, historical operation events and historical operation time of each sample user are obtained in advance, the historical operation events comprise purchase, return, order cancellation, collection cancellation, purchase adding and the like, wherein positive conversion events comprise purchase, collection, purchase adding and the like, and negative conversion behavior events comprise return, order cancellation, collection cancellation and the like. And then marking the positive conversion behavior events in the sample users as positive samples, marking the negative conversion behavior events of the sample users as negative samples, and then training the time sequence diagram neural network by adopting the obtained samples.
And after the training is finished, inputting the behavior sequence into a timing diagram neural network as a graph, acquiring vector representation of the behavior sequence by adopting the timing diagram neural network, and taking the vector representation as a second feature vector of the user to be recommended.
And S403, fusing the first feature vector and the second feature vector to obtain preference features of the user to be recommended for the article.
In specific implementation, the first feature vector and the second feature vector can be fused by adopting a full communication layer, so that the preference feature of the user to be recommended for the article is obtained.
In the embodiment of the application, the attribute information and the behavior sequence of the user to be recommended are respectively extracted and fused by adopting the convolutional neural network and the timing diagram neural network, and the multi-dimensional characteristics of the user to be recommended are obtained, so that the accuracy of the constructed user portrait is improved, and the accuracy of article recommendation is further improved.
Optionally, in the step S402, a behavior sequence diagram may be constructed based on the behavior sequence, and then the behavior sequence diagram is subjected to feature extraction by using a neural network of the behavior sequence diagram, so as to obtain a second feature vector. Specifically, each article and operation event are taken as nodes, edges are determined based on the historical operation events of the user to be recommended for each article in the behavior sequence, and the historical operation time in the behavior sequence is taken as an edge attribute to construct a behavior sequence diagram of the user to be recommended. And extracting the spatial features of the behavior time sequence chart of the user to be recommended by adopting a time sequence chart neural network to obtain a second feature vector of the user to be recommended.
In one possible embodiment, the items corresponding to the nodes in the action sequence chart are all items that can be recommended to the user, such as all goods on sale, all advertisements to be delivered, and the like. The operation events corresponding to the nodes in the action timing diagram are all possible operation events of the user, such as purchase, return, order cancel, collection cancel, purchase adding and the like in a shopping scene, and click, browse, consultation, purchase, return, order cancel, collection cancel, purchase adding and the like in an advertisement putting scene.
In one possible implementation manner, the items corresponding to the nodes in the action timing diagram are all items associated with the user to be recommended, for example, all items associated with the user to be recommended are item a, item B, and item C, and then the items corresponding to the nodes in the action timing diagram are item a, item B, and item C. The operation events corresponding to the nodes in the action timing diagram are all operation events associated with the user to be recommended. For example, if all the operation events associated with the user to be recommended are purchase, return and collection, the operation events corresponding to the nodes in the behavior sequence diagram are purchase, return and collection.
In specific implementation, a time sequence graph neural network is adopted to extract spatial features of a behavior time sequence graph of a user to be recommended, and before a second feature vector of the user to be recommended is obtained, a behavior sequence of each sample user is obtained, wherein the behavior sequence comprises historical operation events and historical operation time of the sample user, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events. And then constructing a sample behavior time sequence chart of the sample users based on the behavior sequence of each sample user, marking positive conversion behavior events in the sample behavior time sequence chart as positive samples, and marking negative conversion behavior events in the sample behavior time sequence chart as negative samples. And then training the time sequence diagram neural network by adopting the sample behavior time sequence chart of each sample user. After training is finished, extracting the spatial features of the behavior time sequence diagram of the user to be recommended by adopting a time sequence diagram neural network, and obtaining a second feature vector of the user to be recommended.
The following illustrates a process of constructing a behavior sequence chart of a user to be recommended.
Illustratively, in the shopping scenario, it is set that the user to be recommended purchases an article a and an article B at time t1, returns the article B to the merchant at time t2, and then collects and purchases an article C at time t 3. The commodities associated with the user to be recommended are a commodity A, a commodity B and a commodity C, the operation events associated with the user to be recommended are purchase, return and collection, and the commodity A, the commodity B, the commodity C, the purchase, the return and the collection are used as nodes of the behavior sequence chart.
As shown in fig. 6a, the behavior sequence diagram corresponding to the t1 time period is that the node "commodity a" is connected with the node "purchase", the node "commodity B" is connected with the node "purchase", and the obtained edge attributes of the two connecting edges are t1 time period.
the behavior sequence diagram corresponding to the time period t2 is shown in fig. 6B, on the basis of fig. 6a, the node "commodity B" is connected with the node "return", and the obtained edge attribute of one connecting edge is the time period t 2.
the behavior sequence diagram corresponding to the time period t3 is shown in fig. 6C, on the basis of fig. 6b, the node "commodity C" is connected with the node "purchase", the node "commodity C" is connected with the node "collection", and the obtained edge attributes of the two connecting edges are the time period t 3.
In a possible implementation manner, when the timing graph neural network is adopted to extract the spatial features of the behavior timing graph of the user to be recommended, all three behavior timing graphs shown in fig. 6a to 6c are input into the timing graph neural network, and the timing graph neural network extracts the spatial features of the three behavior timing graphs to obtain the second feature vector of the user to be recommended.
In another possible implementation manner, when the spatial feature of the behavior time sequence diagram of the user to be recommended is extracted by using the time sequence diagram neural network, the behavior time sequence diagram shown in fig. 6c is input into the time sequence diagram neural network, and the spatial feature of the behavior time sequence diagram is extracted by using the time sequence diagram neural network to obtain the second feature vector of the user to be recommended.
It should be noted that the behavior timing chart of the input timing chart neural network is not limited to the above two embodiments, and other embodiments such as inputting both behavior timing charts shown in fig. 6a and 6c to the timing chart neural network may be used, and the present application is not limited to this embodiment.
Illustratively, in the advertisement placement scenario, it is set that the user to be recommended clicks the advertisement L at the time t1, and then clicks and consults the goods in the advertisement M at the time t 2. Thereafter, at time t3, advertisement N was clicked on and the merchandise in advertisement N was purchased. The advertisements associated with the user to be recommended are advertisement L, advertisement M and advertisement N, and the operation events associated with the user to be recommended are clicking, consulting and purchasing, and then the advertisement L, the advertisement M, the advertisement N, the clicking, the consulting and the purchasing are taken as nodes of the behavior sequence diagram.
the behavior sequence diagram corresponding to the time period t1 is shown in fig. 7a, where the node "ad L" is connected with the node "click", and the obtained edge attribute of one connecting edge is the time period t 1.
the behavior sequence diagram corresponding to the time period t2 is shown in fig. 7b, on the basis of fig. 7a, the node "advertisement M" is connected with the node "consultation", the node "advertisement M" is connected with the node "click", and the obtained edge attribute of the two connecting edges is the time period t 2.
the behavior sequence diagram corresponding to the time period t3 is shown in fig. 7c, on the basis of fig. 7b, the node "ad N" is connected with the node "click", the node "ad N" is connected with the node "purchase", and the obtained edge attribute of one connecting edge is the time period t 3.
In a possible implementation manner, when the timing graph neural network is adopted to extract the spatial features of the behavior timing graph of the user to be recommended, all three behavior timing graphs shown in fig. 7a to 7c are input into the timing graph neural network, and the timing graph neural network extracts the spatial features of the three behavior timing graphs to obtain the second feature vector of the user to be recommended.
In another possible implementation manner, when the spatial feature of the behavior time sequence diagram of the user to be recommended is extracted by using the time sequence diagram neural network, the behavior time sequence diagram shown in fig. 7c is input into the time sequence diagram neural network, and the spatial feature of the behavior time sequence diagram is extracted by using the time sequence diagram neural network to obtain the second feature vector of the user to be recommended.
It should be noted that the behavior timing chart of the input timing chart neural network is not limited to the above two embodiments, and other embodiments such as inputting both behavior timing charts shown in fig. 7a and 7c to the timing chart neural network may be used, and the present application is not particularly limited thereto.
In the embodiment of the application, the behavior sequence diagram of the user to be recommended is constructed based on the behavior sequence of the user to be recommended, then the spatial features of the behavior sequence diagram are extracted by adopting a neural network of the sequence diagram, and the feature vectors on the time sequence of the user to be recommended are obtained, so that the preference features of the user to be recommended, which dynamically change, can be better described, and the accuracy of recommending articles to the user to be recommended is improved.
Optionally, in step S203, the item characteristics of each candidate item may be obtained at least in the following ways:
in a possible implementation manner, the item feature of each candidate item is obtained by performing feature extraction on the description information of each candidate item by using a recurrent neural network.
In specific implementation, the recurrent neural network may be a Bidirectional recurrent neural network (Bi-RNN), a Long-Short-Term Memory network (LSTM), or the like.
The description information of the candidate item can be information such as item category, item price, item brand, item purchase and item comment. For example, if the candidate item is set to be an apple, the description information of the candidate item includes: fruit, 8 yuan each jin, red Fuji, 2000 pieces of monthly sales information and the like.
In a possible implementation manner, the item feature of each candidate item is obtained by performing feature extraction on the description information of each candidate item by using a convolutional neural network.
The convolutional neural network includes at least a convolutional layer, an activation layer, and a pooling layer. And inputting the description information of the candidate object into a convolutional layer, and performing feature extraction on the description information of the candidate object by using convolutional check, wherein the convolutional check in the convolutional layer can be one or more. And inputting the features obtained by convolution into an activation layer, wherein the activation layer adopts an activation function to carry out nonlinear adjustment on the features, and then inputting the features into a pooling layer. And the pooling layer performs down-sampling operation on the adjusted features to obtain the article features of the candidate articles.
The neural network model is adopted to obtain the multi-dimensional object characteristics of the candidate object, so that the characteristics of the candidate object are better represented, the accuracy of subsequent matching with the preference characteristics of the user to be recommended is improved, and the accuracy of object recommendation is improved.
Optionally, in step S204, when recommending an article for the user to be recommended, first matching the preference feature of the user to be recommended for the article with the article feature of each candidate article to obtain a matching weight of each candidate article, and then determining the candidate article with the matching weight greater than a preset threshold as the target article recommended to the user to be recommended.
Specifically, the preference feature of the user to be recommended for the item may be matched with the item feature of each candidate item by using a classification layer, so as to obtain a matching weight of each candidate item. The higher the matching weight is, the more preference the candidate item is currently preferred by the user to be recommended. In one embodiment, the classification layer may be a fully connected layer. The preset threshold may be set according to actual needs, for example, set to 0.8, and when the matching weight of the candidate item is greater than 0.8, determine the candidate item as the target item recommended to the user to be recommended.
The matching weight is obtained by matching the preference characteristics of the user to be recommended for the object with the object characteristics of each candidate object, so that each candidate object is scored, the candidate object with a high score is recommended to the user to be recommended, the user to be recommended can find favorite objects conveniently, and the user experience is improved.
To better explain the embodiment of the present application, an article recommendation method provided by the embodiment of the present application is described below by taking a shopping scenario as an example, where a flow of the method is executed by a recommendation device, which may be the terminal device 101 or the server 102 shown in fig. 1, as shown in fig. 8, and includes the following steps:
and acquiring attribute information of each sample user in advance, wherein the attribute information comprises account registration time, gender, age, location and the like. And obtaining historical operation events and historical operation time of each sample user, wherein the historical operation events comprise purchase, return, order cancellation, collection cancellation, purchase adding and the like, positive conversion events comprise purchase, collection, purchase adding and the like, and negative conversion behavior events comprise return, order cancellation, collection cancellation and the like. And constructing a sample behavior time sequence chart corresponding to each sample based on the historical operation events and the historical operation time of each sample user, marking the positive conversion behavior events in the sample behavior time sequence chart as positive samples, and marking the negative conversion behavior events in the sample behavior time sequence chart as negative samples. The description information of each sample commodity is obtained, and the description information of the sample commodity can be information such as commodity category, commodity price, commodity brand, commodity purchase and commodity comment.
And performing combined training on a recommendation network consisting of a convolutional neural network, a timing sequence neural network and a cyclic neural network by adopting the attribute information of each sample user, the sample behavior timing sequence and the description information of each sample commodity, inputting the attribute information of each sample user into the convolutional neural network, inputting the sample behavior timing sequence of each sample user into the timing sequence neural network and inputting the description information of each sample commodity into the cyclic neural network during training.
After training is finished, acquiring attribute information of a user to be recommended, wherein the attribute information is specifically female, 20 years old and Shanghai, inputting the attribute information of the user to be recommended into a convolutional neural network in a recommendation network, and acquiring a first feature vector of the user to be recommended.
Acquiring historical operation events and historical operation time of a user to be recommended, specifically: the user to be recommended purchased article a and article B at time t1, returned article B to the merchant at time t2, and then collected and purchased article C at time t 3. According to the historical operation events of the user to be recommended, the commodities associated with the user to be recommended are a commodity A, a commodity B and a commodity C, the operation events associated with the user to be recommended are purchase, return and collection, and the commodities A, the commodities B and the commodities C, the purchase, the return and the collection are used as nodes of the behavior sequence diagram. And then constructing a behavior sequence chart of the user to be recommended, specifically, a t1 time period: the node "commodity a" is connected with the node "purchase", the node "commodity B" is connected with the node "purchase", and the obtained edge attributes of the two connecting edges are in a period t 1. Period t 2: on the basis of the time period t1, the node "commodity B" is connected with the node "return goods", and the obtained edge attribute of one connecting edge is the time period t 2. Period t 3: on the basis of the time period t2, the node "commodity C" is connected with the node "purchase", the node "commodity C" is connected with the node "collection", and the obtained edge attributes of the two connecting edges are in the time period t 3. the behavior time charts of the t1 time period, the t2 time period and the t3 time period are input into a time chart neural network in the recommendation network, and a second feature vector of the user to be recommended is obtained.
And inputting the first characteristic vector and the second characteristic vector into a full communication layer in a recommendation network to obtain the preference characteristics of the user to be recommended for the commodity.
And acquiring the description information of each candidate commodity, and inputting the description information of each candidate commodity into a recurrent neural network in the recommendation network to acquire the commodity characteristics of each candidate commodity.
And finally, inputting the preference characteristics of the user to be recommended for the commodities and the commodity characteristics of each candidate commodity into a classification layer in a recommendation network to obtain the matching weight of each candidate commodity, determining the candidate commodities with the matching weights larger than a preset threshold as target commodities recommended to the user to be recommended, and then recommending the target commodities to the user to be recommended.
In the embodiment of the application, as the historical operation events in the behavior sequence comprise positive conversion behavior events and negative conversion behavior events, the positive conversion behavior events are positively correlated with the preference features, and the negative conversion behavior events are negatively correlated with the preference features, when the attribute information of the user to be recommended and the behavior sequence are combined to perform feature extraction on the user, more comprehensive user preference features can be obtained, and then a user image with higher accuracy is constructed. Furthermore, when the articles are recommended to the user according to the user preference characteristics and the article characteristics of the articles, the article recommendation accuracy can be effectively improved.
Based on the same technical concept, an embodiment of the present application provides a schematic structural diagram of a recommendation device, as shown in fig. 9, the recommendation device 900 includes:
an obtaining module 901, configured to obtain attribute information and a behavior sequence of a user to be recommended, where the behavior sequence is obtained by sorting, according to historical operation time, historical operation events of the user to be recommended for each article, where the historical operation events include positive conversion behavior events and negative conversion behavior events;
the processing module 902 is configured to determine, according to the attribute information and the behavior sequence of the user to be recommended, a preference feature of the user to be recommended for the item, where a positive conversion behavior event is positively correlated with the preference feature, and a negative conversion behavior event is negatively correlated with the preference feature;
and the recommending module 903 is configured to determine a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
Optionally, the processing module 902 is specifically configured to:
performing feature extraction on attribute information of a user to be recommended by adopting a convolutional neural network to obtain a first feature vector of the user to be recommended;
performing feature extraction on the behavior sequence by adopting a timing chart neural network to obtain a second feature vector of the user to be recommended, wherein the timing chart neural network is obtained by training by taking a positive conversion behavior event of a sample user as a positive sample and taking a negative conversion behavior event of the sample user as a negative sample;
and fusing the first feature vector and the second feature vector to obtain the preference feature of the user to be recommended for the article.
Optionally, the processing module 902 is specifically configured to:
taking each article and the operation event as nodes, determining edges according to the historical operation events of each article by the user to be recommended in the behavior sequence, and taking the historical operation time in the behavior sequence as edge attributes to construct a behavior sequence diagram of the user to be recommended;
and extracting the spatial features of the behavior time sequence chart of the user to be recommended by adopting a time sequence chart neural network to obtain a second feature vector of the user to be recommended.
Optionally, the article feature of each candidate article is obtained by performing feature extraction on the description information of each candidate article by using a recurrent neural network.
Optionally, the recommending module 903 is specifically configured to:
matching the preference characteristics of the user to be recommended for the articles with the article characteristics of each candidate article to obtain the matching weight of each candidate article;
and determining the candidate object with the matching weight value larger than the preset threshold value as a target object recommended to the user to be recommended.
Based on the same technical concept, the embodiment of the present application provides a computer device, as shown in fig. 10, including at least one processor 1001 and a memory 1002 connected to the at least one processor, where a specific connection medium between the processor 1001 and the memory 1002 is not limited in the embodiment of the present application, and the processor 1001 and the memory 1002 in fig. 10 are connected through a bus as an example. The bus may be divided into an address bus, a data bus, a control bus, etc.
In the embodiment of the present application, the memory 1002 stores instructions executable by the at least one processor 1001, and the at least one processor 1001 may execute the steps included in the aforementioned item recommendation method by executing the instructions stored in the memory 1002.
The processor 1001 is a control center of the computer device, and may connect various parts of the computer device by using various interfaces and lines, and recommend an item to the user to be recommended by executing or executing the instructions stored in the memory 1002 and calling up the data stored in the memory 1002. Alternatively, the processor 1001 may include one or more processing units, and the processor 1001 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, an application program, and the like, and the modem processor mainly processes wireless communication. It will be appreciated that the modem processor described above may not be integrated into the processor 1001. In some embodiments, the processor 1001 and the memory 1002 may be implemented on the same chip, or in some embodiments, they may be implemented separately on separate chips.
The processor 1001 may be a general-purpose processor, such as a Central Processing Unit (CPU), a digital signal processor, an Application Specific Integrated Circuit (ASIC), a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof, and may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present Application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware processor, or may be implemented by a combination of hardware and software modules in a processor.
Memory 1002, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The Memory 1002 may include at least one type of storage medium, and may include, for example, a flash Memory, a hard disk, a multimedia card, a card-type Memory, a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a charge Erasable Programmable Read Only Memory (EEPROM), a magnetic Memory, a magnetic disk, an optical disk, and so on. The memory 1002 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to such. The memory 1002 in the embodiments of the present application may also be circuitry or any other device capable of performing a storage function for storing program instructions and/or data.
Based on the same inventive concept, embodiments of the present application provide a computer-readable storage medium storing a computer program executable by a computer device, which, when the program is run on the computer device, causes the computer device to perform the steps of the above item recommendation method.
It should be apparent to those skilled in the art that embodiments of the present invention may be provided as a method, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (12)

1. An item recommendation method, comprising:
acquiring attribute information and a behavior sequence of a user to be recommended, wherein the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each article according to historical operation time, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events;
determining preference characteristics of the user to be recommended for an article according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics;
and determining a target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
2. The method of claim 1, wherein the determining the preference characteristics of the user to be recommended for the item according to the attribute information and the behavior sequence of the user to be recommended comprises:
performing feature extraction on the attribute information of the user to be recommended by adopting a convolutional neural network to obtain a first feature vector of the user to be recommended;
performing feature extraction on the behavior sequence by adopting a time sequence neural network to obtain a second feature vector of the user to be recommended, wherein the time sequence neural network is obtained by training by taking a positive conversion behavior event of a sample user as a positive sample and taking a negative conversion behavior event of the sample user as a negative sample;
and fusing the first feature vector and the second feature vector to obtain the preference feature of the user to be recommended for the article.
3. The method of claim 2, wherein the performing feature extraction on the behavior sequence by using a timing graph neural network to obtain a second feature vector of the user to be recommended comprises:
taking each article and the operation event as nodes, determining edges based on the historical operation events of the to-be-recommended user for each article in the behavior sequence, and constructing a behavior sequence diagram of the to-be-recommended user by taking the historical operation time in the behavior sequence as edge attributes;
and extracting the spatial features of the behavior time sequence diagram of the user to be recommended by adopting a time sequence diagram neural network to obtain a second feature vector of the user to be recommended.
4. The method of claim 1, wherein the item features of the candidate items are obtained by feature extraction of description information of the candidate items using a recurrent neural network.
5. The method according to any one of claims 1 to 4, wherein the determining of the target item recommended to the user to be recommended from each candidate item according to the preference feature of the user to be recommended for the item and the item feature of each candidate item comprises:
matching the preference characteristics of the user to be recommended for the articles with the article characteristics of each candidate article to obtain the matching weight of each candidate article;
and determining the candidate object with the matching weight value larger than a preset threshold value as a target object recommended to the user to be recommended.
6. A recommendation device, comprising:
the system comprises an acquisition module, a recommendation module and a recommendation module, wherein the acquisition module is used for acquiring attribute information and a behavior sequence of a user to be recommended, the behavior sequence is obtained by sequencing historical operation events of the user to be recommended aiming at each article according to historical operation time, and the historical operation events comprise positive conversion behavior events and negative conversion behavior events;
the processing module is used for determining preference characteristics of the user to be recommended for the article according to the attribute information and the behavior sequence of the user to be recommended, wherein the positive conversion behavior event is positively correlated with the preference characteristics, and the negative conversion behavior event is negatively correlated with the preference characteristics;
and the recommending module is used for determining the target item recommended to the user to be recommended from each candidate item according to the preference characteristics of the user to be recommended for the item and the item characteristics of each candidate item.
7. The recommendation device of claim 6, wherein the processing module is specifically configured to:
performing feature extraction on the attribute information of the user to be recommended by adopting a convolutional neural network to obtain a first feature vector of the user to be recommended;
performing feature extraction on the behavior sequence by adopting a time sequence neural network to obtain a second feature vector of the user to be recommended, wherein the time sequence neural network is obtained by training by taking a positive conversion behavior event of a sample user as a positive sample and taking a negative conversion behavior event of the sample user as a negative sample;
and fusing the first feature vector and the second feature vector to obtain the preference feature of the user to be recommended for the article.
8. The recommendation device of claim 7, wherein the processing module is specifically configured to:
taking each article and the operation event as nodes, determining edges based on the historical operation events of the to-be-recommended user for each article in the behavior sequence, and constructing a behavior sequence diagram of the to-be-recommended user by taking the historical operation time in the behavior sequence as edge attributes;
and extracting the spatial features of the behavior time sequence diagram of the user to be recommended by adopting a time sequence diagram neural network to obtain a second feature vector of the user to be recommended.
9. The recommendation device of claim 6, wherein the item features of the respective candidate items are obtained by feature extraction of description information of the respective candidate items using a recurrent neural network.
10. The recommendation device of any of claims 6 to 9, wherein the recommendation module is specifically configured to:
matching the preference characteristics of the user to be recommended for the articles with the article characteristics of each candidate article to obtain the matching weight of each candidate article;
and determining the candidate object with the matching weight value larger than a preset threshold value as a target object recommended to the user to be recommended.
11. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the method of any one of claims 1 to 5 are performed when the program is executed by the processor.
12. A computer-readable storage medium, having stored thereon a computer program executable by a computer device, for causing the computer device to perform the steps of the method of any one of claims 1 to 5, when the program is run on the computer device.
CN202010615163.6A 2020-06-30 2020-06-30 Article recommendation method and recommendation equipment Active CN111784455B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010615163.6A CN111784455B (en) 2020-06-30 2020-06-30 Article recommendation method and recommendation equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010615163.6A CN111784455B (en) 2020-06-30 2020-06-30 Article recommendation method and recommendation equipment

Publications (2)

Publication Number Publication Date
CN111784455A true CN111784455A (en) 2020-10-16
CN111784455B CN111784455B (en) 2023-11-07

Family

ID=72760875

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010615163.6A Active CN111784455B (en) 2020-06-30 2020-06-30 Article recommendation method and recommendation equipment

Country Status (1)

Country Link
CN (1) CN111784455B (en)

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112256545A (en) * 2020-10-21 2021-01-22 武汉悦学帮网络技术有限公司 Method and device for acquiring user operation information
CN112330407A (en) * 2020-11-11 2021-02-05 北京沃东天骏信息技术有限公司 Article matching method and device and method for acquiring article matching model
CN112365062A (en) * 2020-11-13 2021-02-12 中国民航大学 Method and system for mining implicit interactive features and recommending flights of civil aviation passengers
CN112801191A (en) * 2021-02-02 2021-05-14 中国石油大学(北京) Intelligent recommendation method, device and equipment for pipeline accident handling
CN112950321A (en) * 2021-03-10 2021-06-11 北京汇钧科技有限公司 Article recommendation method and device
CN112991026A (en) * 2021-05-08 2021-06-18 明品云(北京)数据科技有限公司 Commodity recommendation method, system, equipment and computer readable storage medium
CN113077317A (en) * 2021-04-19 2021-07-06 北京京东拓先科技有限公司 Item recommendation method, device and equipment based on user data and storage medium
CN113128995A (en) * 2021-05-18 2021-07-16 中国银行股份有限公司 Two-dimensional code recommendation method and device based on mobile banking
CN113158057A (en) * 2021-04-28 2021-07-23 平安科技(深圳)有限公司 Buddha meridian recommendation processing device, computer equipment and storage medium
CN113239091A (en) * 2021-05-14 2021-08-10 杭州志卓科技股份有限公司 Intelligent evaluation system for artificial intelligence B2B website users
CN113377849A (en) * 2021-05-24 2021-09-10 上海钧正网络科技有限公司 Information recommendation method, device and equipment
CN113435969A (en) * 2021-06-23 2021-09-24 未鲲(上海)科技服务有限公司 Product recommendation method and device, electronic equipment and storage medium
CN113469752A (en) * 2021-07-22 2021-10-01 北京沃东天骏信息技术有限公司 Content recommendation method and device, storage medium and electronic equipment
CN113486235A (en) * 2021-05-31 2021-10-08 杭州摸象大数据科技有限公司 Method and system for identifying user interest
CN113723733A (en) * 2020-12-23 2021-11-30 北京沃东天骏信息技术有限公司 Article information display method and device, electronic equipment and computer readable medium
CN113781147A (en) * 2020-11-17 2021-12-10 北京沃东天骏信息技术有限公司 Item recommendation method and device
CN113836407A (en) * 2021-09-14 2021-12-24 马上消费金融股份有限公司 Recommendation method and related device
CN113902028A (en) * 2021-10-22 2022-01-07 北京达佳互联信息技术有限公司 Model training method for category identification, category identification method and related device
CN114187062A (en) * 2021-11-10 2022-03-15 深圳童尔家教育咨询有限公司 Commodity purchase event prediction method and device
WO2023284516A1 (en) * 2021-07-16 2023-01-19 腾讯科技(深圳)有限公司 Information recommendation method and apparatus based on knowledge graph, and device, medium, and product
CN116628345A (en) * 2023-07-13 2023-08-22 腾讯科技(深圳)有限公司 Content recommendation method and device, electronic equipment and storage medium

Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180103052A1 (en) * 2016-10-11 2018-04-12 Battelle Memorial Institute System and methods for automated detection, reasoning and recommendations for resilient cyber systems
US10262079B1 (en) * 2018-09-25 2019-04-16 Accenture Global Solutions Limited Determining anonymized temporal activity signatures of individuals
US20190311367A1 (en) * 2015-06-20 2019-10-10 Quantiply Corporation System and method for using a data genome to identify suspicious financial transactions
US20190312898A1 (en) * 2018-04-10 2019-10-10 Cisco Technology, Inc. SPATIO-TEMPORAL ANOMALY DETECTION IN COMPUTER NETWORKS USING GRAPH CONVOLUTIONAL RECURRENT NEURAL NETWORKS (GCRNNs)
CN110334245A (en) * 2019-05-20 2019-10-15 山东大学 A kind of short video recommendation method and device of the figure neural network based on Temporal Order
CN110413877A (en) * 2019-07-02 2019-11-05 阿里巴巴集团控股有限公司 A kind of resource recommendation method, device and electronic equipment
CN110427560A (en) * 2019-08-08 2019-11-08 腾讯科技(深圳)有限公司 A kind of model training method and relevant apparatus applied to recommender system
CN110442802A (en) * 2019-08-06 2019-11-12 中国科学技术大学 A kind of more Behavior preference prediction techniques of social activity user
CN110490717A (en) * 2019-09-05 2019-11-22 齐鲁工业大学 Method of Commodity Recommendation and system based on user conversation and figure convolutional neural networks
CN110489639A (en) * 2019-07-15 2019-11-22 北京奇艺世纪科技有限公司 A kind of content recommendation method and device
CN110598084A (en) * 2018-05-24 2019-12-20 阿里巴巴集团控股有限公司 Object sorting method, commodity sorting device and electronic equipment
KR102091529B1 (en) * 2019-09-03 2020-03-23 (주)빅인사이트 Method and apparatus for training AI model using user's time series behavior data
CN110910209A (en) * 2019-11-12 2020-03-24 腾讯科技(深圳)有限公司 Data processing method and device and computer readable storage medium
CN110969516A (en) * 2019-12-25 2020-04-07 清华大学 Commodity recommendation method and device
CN111177575A (en) * 2020-04-07 2020-05-19 腾讯科技(深圳)有限公司 Content recommendation method and device, electronic equipment and storage medium
CN111325340A (en) * 2020-02-17 2020-06-23 南方科技大学 Information network relation prediction method and system

Patent Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190311367A1 (en) * 2015-06-20 2019-10-10 Quantiply Corporation System and method for using a data genome to identify suspicious financial transactions
US20180103052A1 (en) * 2016-10-11 2018-04-12 Battelle Memorial Institute System and methods for automated detection, reasoning and recommendations for resilient cyber systems
US20190312898A1 (en) * 2018-04-10 2019-10-10 Cisco Technology, Inc. SPATIO-TEMPORAL ANOMALY DETECTION IN COMPUTER NETWORKS USING GRAPH CONVOLUTIONAL RECURRENT NEURAL NETWORKS (GCRNNs)
CN110598084A (en) * 2018-05-24 2019-12-20 阿里巴巴集团控股有限公司 Object sorting method, commodity sorting device and electronic equipment
US10262079B1 (en) * 2018-09-25 2019-04-16 Accenture Global Solutions Limited Determining anonymized temporal activity signatures of individuals
CN110334245A (en) * 2019-05-20 2019-10-15 山东大学 A kind of short video recommendation method and device of the figure neural network based on Temporal Order
CN110413877A (en) * 2019-07-02 2019-11-05 阿里巴巴集团控股有限公司 A kind of resource recommendation method, device and electronic equipment
CN110489639A (en) * 2019-07-15 2019-11-22 北京奇艺世纪科技有限公司 A kind of content recommendation method and device
CN110442802A (en) * 2019-08-06 2019-11-12 中国科学技术大学 A kind of more Behavior preference prediction techniques of social activity user
CN110427560A (en) * 2019-08-08 2019-11-08 腾讯科技(深圳)有限公司 A kind of model training method and relevant apparatus applied to recommender system
KR102091529B1 (en) * 2019-09-03 2020-03-23 (주)빅인사이트 Method and apparatus for training AI model using user's time series behavior data
CN110490717A (en) * 2019-09-05 2019-11-22 齐鲁工业大学 Method of Commodity Recommendation and system based on user conversation and figure convolutional neural networks
CN110910209A (en) * 2019-11-12 2020-03-24 腾讯科技(深圳)有限公司 Data processing method and device and computer readable storage medium
CN110969516A (en) * 2019-12-25 2020-04-07 清华大学 Commodity recommendation method and device
CN111325340A (en) * 2020-02-17 2020-06-23 南方科技大学 Information network relation prediction method and system
CN111177575A (en) * 2020-04-07 2020-05-19 腾讯科技(深圳)有限公司 Content recommendation method and device, electronic equipment and storage medium

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
JINGHUA ZHU: "Deep Neural Model for Point-of-Interest Recommendation Fused with Graph Embedding Representation", 《WIRELESS ALGORITHMS, SYSTEMS, AND APPLICATIONS》, pages 495 - 506 *
吴国栋: "图神经网络推荐研究进展", 智能系统学报, vol. 15, no. 01, pages 14 - 24 *
张志扬: "基于分层注意力的信息级联预测模型", 计算机科学, vol. 47, no. 06, pages 201 - 209 *

Cited By (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112256545A (en) * 2020-10-21 2021-01-22 武汉悦学帮网络技术有限公司 Method and device for acquiring user operation information
CN112330407A (en) * 2020-11-11 2021-02-05 北京沃东天骏信息技术有限公司 Article matching method and device and method for acquiring article matching model
CN112365062A (en) * 2020-11-13 2021-02-12 中国民航大学 Method and system for mining implicit interactive features and recommending flights of civil aviation passengers
CN113781147A (en) * 2020-11-17 2021-12-10 北京沃东天骏信息技术有限公司 Item recommendation method and device
CN113723733A (en) * 2020-12-23 2021-11-30 北京沃东天骏信息技术有限公司 Article information display method and device, electronic equipment and computer readable medium
CN112801191A (en) * 2021-02-02 2021-05-14 中国石油大学(北京) Intelligent recommendation method, device and equipment for pipeline accident handling
CN112801191B (en) * 2021-02-02 2023-11-21 中国石油大学(北京) Intelligent recommendation method, device and equipment for handling pipeline accidents
CN112950321A (en) * 2021-03-10 2021-06-11 北京汇钧科技有限公司 Article recommendation method and device
CN113077317A (en) * 2021-04-19 2021-07-06 北京京东拓先科技有限公司 Item recommendation method, device and equipment based on user data and storage medium
CN113158057A (en) * 2021-04-28 2021-07-23 平安科技(深圳)有限公司 Buddha meridian recommendation processing device, computer equipment and storage medium
CN112991026A (en) * 2021-05-08 2021-06-18 明品云(北京)数据科技有限公司 Commodity recommendation method, system, equipment and computer readable storage medium
CN113239091A (en) * 2021-05-14 2021-08-10 杭州志卓科技股份有限公司 Intelligent evaluation system for artificial intelligence B2B website users
CN113128995A (en) * 2021-05-18 2021-07-16 中国银行股份有限公司 Two-dimensional code recommendation method and device based on mobile banking
CN113377849A (en) * 2021-05-24 2021-09-10 上海钧正网络科技有限公司 Information recommendation method, device and equipment
CN113486235A (en) * 2021-05-31 2021-10-08 杭州摸象大数据科技有限公司 Method and system for identifying user interest
CN113435969A (en) * 2021-06-23 2021-09-24 未鲲(上海)科技服务有限公司 Product recommendation method and device, electronic equipment and storage medium
WO2022267160A1 (en) * 2021-06-23 2022-12-29 未鲲(上海)科技服务有限公司 Product recommendation method and apparatus, electronic device and storage medium
WO2023284516A1 (en) * 2021-07-16 2023-01-19 腾讯科技(深圳)有限公司 Information recommendation method and apparatus based on knowledge graph, and device, medium, and product
CN113469752A (en) * 2021-07-22 2021-10-01 北京沃东天骏信息技术有限公司 Content recommendation method and device, storage medium and electronic equipment
CN113836407B (en) * 2021-09-14 2022-08-19 马上消费金融股份有限公司 Recommendation method and related device
CN113836407A (en) * 2021-09-14 2021-12-24 马上消费金融股份有限公司 Recommendation method and related device
CN113902028A (en) * 2021-10-22 2022-01-07 北京达佳互联信息技术有限公司 Model training method for category identification, category identification method and related device
CN114187062A (en) * 2021-11-10 2022-03-15 深圳童尔家教育咨询有限公司 Commodity purchase event prediction method and device
CN114187062B (en) * 2021-11-10 2024-02-20 深圳童尔家教育咨询有限公司 Commodity purchase event prediction method and device
CN116628345A (en) * 2023-07-13 2023-08-22 腾讯科技(深圳)有限公司 Content recommendation method and device, electronic equipment and storage medium
CN116628345B (en) * 2023-07-13 2024-02-06 腾讯科技(深圳)有限公司 Content recommendation method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN111784455B (en) 2023-11-07

Similar Documents

Publication Publication Date Title
CN111784455B (en) Article recommendation method and recommendation equipment
CN110909176B (en) Data recommendation method and device, computer equipment and storage medium
CN109299994B (en) Recommendation method, device, equipment and readable storage medium
CN111898031B (en) Method and device for obtaining user portrait
CN109934619A (en) User's portrait tag modeling method, apparatus, electronic equipment and readable storage medium storing program for executing
WO2023011382A1 (en) Recommendation method, recommendation model training method, and related product
CN111400613A (en) Article recommendation method, device, medium and computer equipment
Shen et al. A voice of the customer real-time strategy: An integrated quality function deployment approach
CN111967924A (en) Commodity recommendation method, commodity recommendation device, computer device, and medium
CN113781149B (en) Information recommendation method and device, computer readable storage medium and electronic equipment
CN111310035A (en) Recommendation method and device based on psychological and behavioral characteristics
CN116764236A (en) Game prop recommending method, game prop recommending device, computer equipment and storage medium
CN118193806A (en) Target retrieval method, target retrieval device, electronic equipment and storage medium
Zhang et al. Garment recommendation in an e-shopping environment by using a Markov Chain and Complex Network integrated method
CN116821516B (en) Resource recommendation method, device, equipment and storage medium
CN115641179A (en) Information pushing method and device and electronic equipment
CN116823410A (en) Data processing method, object processing method, recommending method and computing device
CN110781399A (en) Cross-platform information pushing method and device
Pradeep et al. Data science and deep learning applications in the e-commerce industry: A survey
Nedjah et al. Client profile prediction using convolutional neural networks for efficient recommendation systems in the context of smart factories
CN114429384A (en) Intelligent product recommendation method and system based on e-commerce platform
CN113724044A (en) User portrait based commodity recommendation, apparatus, computer device and storage medium
CN113254775A (en) Credit card product recommendation method based on client browsing behavior sequence
CN112488355A (en) Method and device for predicting user rating based on graph neural network
CN111310028A (en) Recommendation method and device based on psychological characteristics

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40030088

Country of ref document: HK

SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant