CN117407598A - Live broadcast recommendation method and system - Google Patents
Live broadcast recommendation method and system Download PDFInfo
- Publication number
- CN117407598A CN117407598A CN202311321592.2A CN202311321592A CN117407598A CN 117407598 A CN117407598 A CN 117407598A CN 202311321592 A CN202311321592 A CN 202311321592A CN 117407598 A CN117407598 A CN 117407598A
- Authority
- CN
- China
- Prior art keywords
- user
- live broadcast
- commodity
- live
- interest classification
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000013145 classification model Methods 0.000 claims abstract description 24
- 230000009191 jumping Effects 0.000 claims abstract description 17
- 230000006399 behavior Effects 0.000 claims description 74
- 238000004458 analytical method Methods 0.000 claims description 33
- 230000003993 interaction Effects 0.000 claims description 22
- 238000003058 natural language processing Methods 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 12
- 235000019633 pungent taste Nutrition 0.000 claims description 11
- 238000005516 engineering process Methods 0.000 claims description 9
- 230000002452 interceptive effect Effects 0.000 claims description 9
- 230000008451 emotion Effects 0.000 claims description 8
- 238000012163 sequencing technique Methods 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 description 16
- 238000011156 evaluation Methods 0.000 description 11
- 244000025254 Cannabis sativa Species 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 238000001914 filtration Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000005065 mining Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000011426 transformation method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000008094 contradictory effect Effects 0.000 description 1
- 239000002537 cosmetic Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 244000144972 livestock Species 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a live broadcast recommendation method, which comprises the following steps: determining interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform; analyzing the commodity type corresponding to each live broadcast according to live broadcast with goods forecast of each live broadcast; matching the user interest classification with the commodity type of each live broadcast; and recommending live broadcast with-cargo site times to the user according to the matching result. The application also discloses a live broadcast recommendation system, an electronic device and a computer readable storage medium. Therefore, live broadcast with-cargo field recommendation can be performed on commodities possibly interested by the user, and user experience is improved.
Description
Technical Field
The present disclosure relates to the field of live broadcast technologies, and in particular, to a live broadcast recommendation method, a live broadcast recommendation system, an electronic device, and a computer readable storage medium.
Background
Live broadcast with goods is carried by the live broadcast entertainment industry while live broadcast, and commodity recommendation is carried out in a live broadcast room by a host, which is also called as a good recommendation officer. Currently, each video platform and each electronic commerce platform have a large number of live broadcast with cargo orders. However, in the live-stock area, there is no perception of the classification of goods that may be of interest to the user, and no further user recommendation.
Disclosure of Invention
The main objective of the present application is to provide a live broadcast recommendation method, a system, an electronic device and a computer readable storage medium, which aim to solve the problem of how to recommend live broadcast with goods for goods that a user may be interested in.
In order to achieve the above objective, an embodiment of the present application provides a live broadcast recommendation method, where the method includes:
determining interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform;
analyzing the commodity type corresponding to each live broadcast according to live broadcast with goods forecast of each live broadcast;
matching the user interest classification with the commodity type of each live broadcast;
and recommending live broadcast with-cargo site times to the user according to the matching result.
Optionally, before matching the user interest classification with the commodity type of each live broadcast, the method further comprises:
and determining the user interest classification based on the behavior analysis model according to the interactive behavior of the user watching the video on the current platform.
Optionally, the determining the interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform comprises:
acquiring commodity characteristic information corresponding to the commodity browsing record and the jump record;
converting the commodity browsing records and the jump records into behavior feature vectors;
and inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model, and predicting the interest degree of the user on different commodity classifications.
Optionally, the analyzing the commodity type corresponding to each live broadcast according to the live broadcast with the goods forecast of each live broadcast comprises:
and carrying out text analysis and content understanding on live broadcast with goods forenotice of each occasion through a natural language processing technology, and determining the commodity type corresponding to each live broadcast.
Optionally, the determining the user interest classification based on the behavior analysis model according to the interactive behavior of the user watching the video on the current platform includes:
acquiring a barrage or comment sent by a user when watching a video;
extracting keywords in the barrage or the comments through text classification;
carrying out emotion tendency analysis according to the keywords, and determining preference degrees of users on different video contents;
and determining the interest classification of the user according to the commodity information corresponding to the video content.
Optionally, the determining the user interest classification based on the behavior analysis model according to the interactive behavior of the user watching the video on the current platform further comprises:
the method comprises the steps of obtaining interaction behaviors of a user when watching video, wherein the interaction behaviors comprise bullet screen sending, comment sending, praying, coin inserting, forwarding, collecting and sharing behaviors;
determining the preference degree of the user for different video contents according to the weights of various interaction behaviors;
and determining the interest classification of the user according to the commodity information corresponding to the video content.
Optionally, the recommending the live broadcast with the goods yard according to the matching result includes:
sequencing the predicted live broadcast with the goods orders according to the matching degree, and recommending N live broadcast with the goods orders which are ranked forward to the user, wherein N is a preset positive integer.
Optionally, the recommending the live broadcast with the goods yard according to the matching result to the user further includes:
acquiring playing information of each live broadcast, wherein the playing information comprises hotness of a live broadcast room and a live broadcast time period;
and recommending live broadcast with goods yard times to the user by combining the matching result and the playing information.
In addition, in order to achieve the above object, an embodiment of the present application further provides a live broadcast recommendation system, where the system includes:
the determining module is used for determining interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform;
the analysis module is used for analyzing the commodity type corresponding to each live broadcast according to live broadcast with goods forecast of each live broadcast;
the matching module is used for matching the user interest classification with the commodity type of each live broadcast;
and the recommendation module is used for recommending live broadcast on-site times to the user according to the matching result.
To achieve the above object, an embodiment of the present application further provides an electronic device, including: the live broadcast recommendation method comprises a memory, a processor and a live broadcast recommendation program which is stored in the memory and can run on the processor, wherein the live broadcast recommendation program realizes the live broadcast recommendation method when being executed by the processor.
To achieve the above object, an embodiment of the present application further provides a computer readable storage medium, where a live broadcast recommendation program is stored, where the live broadcast recommendation program implements the live broadcast recommendation method described above when executed by a processor.
According to the live broadcast recommending method, system, electronic device and computer readable storage medium, according to the commodity browsing record of the live broadcast with goods and the record of jumping to the third party e-commerce platform and the interaction behavior of watching videos, which commodity classification is interested in the user recently is deduced, then a plurality of live broadcast with goods occasions to be started are matched through the classification, the live broadcast with goods is recommended to the user for reservation, accuracy and pertinence of live broadcast with goods recommendation are improved, and user experience is improved.
Drawings
FIG. 1 is a diagram of an application environment architecture for implementing various embodiments of the present application;
fig. 2 is a flowchart of a live broadcast recommendation method according to a first embodiment of the present application;
FIG. 3 is a schematic diagram of a refinement process of step S200 in FIG. 2;
fig. 4 is a flowchart of a live broadcast recommendation method according to a second embodiment of the present application;
FIG. 5 is a schematic diagram of a refinement procedure of step S300 in FIG. 4;
FIG. 6 is a schematic diagram of another refinement procedure of step S300 in FIG. 4;
FIG. 7 is a schematic flow chart of an embodiment of the present application;
fig. 8 is a schematic hardware architecture of an electronic device according to a third embodiment of the present application;
fig. 9 is a schematic block diagram of a live broadcast recommendation system according to a fourth embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
It should be noted that the descriptions of "first," "second," etc. in the embodiments of the present application are for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions of the embodiments may be combined with each other, but it is necessary to base that the technical solutions can be realized by those skilled in the art, and when the technical solutions are contradictory or cannot be realized, the combination of the technical solutions should be regarded as not exist and not within the protection scope of the present application.
The following provides an explanation of terms involved in the present application:
recommendation algorithm: an algorithm in the computer profession uses some behaviors of the user to infer what the user may like through some mathematical algorithms. Recommendation algorithms are broadly classified into categories of content-based, collaborative filtering-based, association rule-based, utility-based, knowledge-based, and combined recommendation, and the embodiment of the present application is mainly a recommendation algorithm based on association rules.
Carrying: a network term indicates that a star, a network red, an Internet marketer and the like promote goods through shooting videos or live broadcasting and other approaches.
And (3) grass planting: refers to sharing some products or services which are experienced by themselves, and is beneficial to other people through actual measurement of the products or services.
Evaluation: the evaluation and comparison of various commodities of the same type are performed through various aspects, such as evaluation from various aspects of brands, prices, materials/components, functions, use effects and the like of the commodities.
Cargo video, grass planting video, and evaluation video: commodity carrying, grass planting or evaluation is carried out through the medium of video, and consumers are stimulated to generate purchasing desire through lens expression.
Machine learning: is a multi-domain interdisciplinary, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. It is specially studied how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, and reorganizes existing knowledge structures to continuously improve own performance.
Natural language processing (Natural Language Processing, NLP): the method takes the language as an object, utilizes the computer technology to analyze, understand and process a subject of the natural language, namely takes the computer as a powerful tool for language research, quantitatively researches the language information under the support of the computer, and provides language description which can be used together between people and the computer. Natural language processing includes two parts, natural language understanding (NaturalLanguage Understanding, NLU) and Natural language generation (Natural LanguageGeneration, NLG).
Bag of words model (Bag-of-words model): is a simplified expression model under natural language processing and information retrieval. Under the bag-of-words model, words such as sentences or documents can be expressed in a bag that holds the words regardless of grammar and word order. The bag of words model is widely used for document classification, and the frequency of word occurrence can be used as a feature for training a classifier.
TF-IDF (Term Frequency-inverse text Frequency index): a common weighting technique for information retrieval and data mining is used to evaluate the importance of a word to one of a set of documents or a corpus. The importance of a word increases proportionally with the number of times it appears in the file, but at the same time decreases inversely with the frequency with which it appears in the corpus.
Referring to fig. 1, fig. 1 is a schematic diagram of an application environment for implementing various embodiments of the present application. The application is applicable to application environments including, but not limited to, client 2, server 4, and network 6.
The client side 2 is used for receiving operations such as commodity browsing, third-party e-commerce platform skipping, barrage sending and the like of a user, and displaying recommended live broadcast with cargo orders to the user. The client 2 may be a terminal device such as a PC (Personal Computer ), a mobile phone, a tablet computer, a portable computer, or a wearable device.
The server 4 is configured to provide data and technical support for the client 2, for example, to analyze user interest classification, match live broadcast with goods orders of interest for users, and the like. The server 4 may be a rack server, a blade server, a tower server, or a cabinet server, and may be an independent server or a server cluster formed by a plurality of servers.
The network 6 may be a wireless or wired network such as an Intranet (Intranet), the Internet (Internet), a global system for mobile communications (Global System of Mobile communication, GSM), wideband code division multiple access (Wideband Code Division Multiple Access, WCDMA), a 4G network, a 5G network, bluetooth (Bluetooth), wi-Fi, etc. The server 4 and one or more clients 2 are in communication connection through the network 6 for data transmission and interaction.
Example 1
Fig. 2 is a flowchart of a live broadcast recommendation method according to a first embodiment of the present application. It will be appreciated that the flow charts in the method embodiments are not intended to limit the order in which the steps are performed. Some of the steps in the flow chart may be added or subtracted as desired. The method will be described below with the server as an execution subject.
The method comprises the following steps:
and S200, determining the interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform.
By using big data in the background, a large amount of user behavior data can be acquired and analyzed. Through the behavior data, user portraits can be established and analyzed, and information such as interest preferences and consumption habits of users can be known. In terms of large data processing, a large data set may be processed using a distributed computing framework (e.g., hadoop, spark) to quickly extract and analyze user behavior data. Meanwhile, the data mining and machine learning algorithm can be utilized to deeply mine the user behavior data, and a hidden association rule and a hidden user behavior mode can be found.
In this embodiment, a commodity browsing record of the live broadcasting room with the current platform and a record of jumping to a third party e-commerce platform in a preset period (for example, one week) of the user are obtained. For example, the user views live broadcast with goods in a live broadcast room of the current platform, clicks on goods in the goods list, opens a goods detail page, and successfully jumps from the goods detail page to a third party e-commerce platform such as Taobao, jindong, spell-up-multiple. From these behavior records, the interest classification model may be constructed, analyzing the user's interest classification.
Specifically, referring further to fig. 3, a schematic refinement flow chart of the above step S200 is shown. It will be appreciated that the flowchart is not intended to limit the order in which the steps are performed. Some of the steps in the flow chart may be added or subtracted as desired. In this embodiment, the step S200 specifically includes:
and S2000, acquiring commodity characteristic information corresponding to the commodity browsing record and the skip record.
And according to the commodity browsing record and the jump record, corresponding commodity characteristic information can be obtained. For example, aiming at the behavior record of a user opening a certain commodity detail page, commodity information corresponding to the commodity detail page is obtained. The commodity characteristic information includes commodity names, brands, prices, and the like.
S2002, converting the commodity browsing record and the jump record into behavior feature vectors.
And converting the behavior of the user into a feature vector according to the commodity browsing record and the jump record. The specific transformation method can adopt the existing eigenvector transformation algorithm, and is not limited herein.
And S2004, inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model, and predicting the interest degree of the user on different commodity classifications.
The interest classification model may use machine learning algorithms such as decision trees, random forests, support vector machines, and the like. The interest degree of the user for classifying different commodities can be predicted by training the interest classification model. Inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model can output a prediction result of interest classification of the user, namely, what commodity types (such as home furnishing and cosmetic) the user is interested in, and the degree of interest is.
Returning to fig. 2, S202 analyzes the commodity type corresponding to each live broadcast according to the live broadcast with the forecast of each live broadcast.
In this embodiment, text analysis and content understanding can be performed on live broadcast and live broadcast previews of each site broadcast by a live broadcast owner through a natural language processing technology, and a commodity type corresponding to each site of live broadcast is determined.
S204, matching the user interest classification with the commodity type of each live broadcast.
And judging the correlation between the commodity type of each live broadcast and the interest classification of the user for the live broadcast with the cargo scene to be broadcast. Based on the commodity classification matching model, text mining techniques, such as word bag models, TF-IDF, etc., may be used to match the user interest classification with the commodity type of each live broadcast. Specifically, the description text of the commodity may be represented in a vectorization manner, and then a similarity (matching degree) between the user interest classification and the commodity type is calculated, so that the commodity type related to the user interest is found.
S206, recommending live broadcast on-demand occasions to the user according to the matching result.
And sequencing the forecast live broadcast with the cargo field according to the matching degree obtained in the last step. Therefore, the N live broadcast with the top-ranked orders can be recommended to the user, wherein N is a preset positive integer. In other embodiments, a matching degree threshold may be set, and then the live-broadcast on-demand time with the calculated matching degree reaching the threshold may be recommended to the user.
In an alternative embodiment, in addition to considering the user interest classification, other factors such as the hotness of the live room, whether the live time period user is idle, etc. need to be considered in recommending to the user. Specifically, firstly, the playing information of each live broadcast is obtained, wherein the playing information comprises the hotness of a live broadcast room, the live broadcast time period and the like. And then, recommending live broadcast with the goods yard to the user by combining the matching result and the playing information. For example, live orders with high hotness in the live rooms are recommended preferentially in addition to the ranking according to the matching degree. For another example, according to the historical behavior data of the user, judging an idle time period of the live broadcast watched by the user frequently, and recommending the live broadcast time which is matched with the idle time period with priority. In practical application, corresponding weights and formulas can be formulated according to service requirements.
According to the live broadcast recommendation method, according to the commodity browsing records of the user in the live broadcast room and the records of the third-party electronic commerce platform, the user can infer which commodity classification is interested in recently, and then the classification is matched with a plurality of live broadcast with goods occasions to be broadcast, so that the live broadcast with goods is recommended to the user for reservation, the accuracy and pertinence of live broadcast with goods recommendation are improved, and the user experience is improved.
Example two
Fig. 4 is a flowchart of a live broadcast recommendation method according to a second embodiment of the present application. In a second embodiment, the live broadcast recommendation method further increases analysis of interaction behavior of a user watching video on the basis of the first embodiment. It will be appreciated that the flow charts in the method embodiments are not intended to limit the order in which the steps are performed. Some of the steps in the flow chart may be added or subtracted as desired.
The method comprises the following steps:
s300, determining interest classification of the user based on the interest classification model and the behavior analysis model according to the commodity browsing record of the user on the current platform, the record of jumping to the third-party e-commerce platform and the interaction behavior of the user watching the video on the current platform.
In this embodiment, the user interest classification may be determined by both the user's merchandise browsing record on the current platform and the jump to the third party e-commerce platform record, and the user's interaction behavior in viewing video on the current platform. First, a commodity browsing record of a live broadcast room with a current platform and a record of jumping to a third party e-commerce platform in a preset time period (for example, one week) of a user are obtained. For example, the user views live broadcast with goods in a live broadcast room of the current platform, clicks on goods in the goods list, opens a goods detail page, and successfully jumps from the goods detail page to a third party e-commerce platform such as Taobao, jindong, spell-up-multiple. From these behavior records, the interest classification model may be constructed, analyzing the user's interest classification. The implementation principle of determining the user interest classification based on the interest classification model is the same as that of steps S2000-S2004 in the foregoing first embodiment, and the specific implementation process may refer to the description in the first embodiment, which is not repeated herein.
In addition, in the preset time period, the interaction behavior of the user when the user watches the grass planting video, the goods carrying video and the evaluation video on the current platform can be obtained, such as the behavior of bullet screen sending, comment, praise, coin-in, forwarding, collection, sharing and the like. From these above behaviors, a user interest classification may also be determined based on the behavior analysis model.
In an alternative embodiment, the viewing behavior analysis model may analyze the bullet screen or comment generated by the user while viewing the grass video, the cargo video, the review video on the current platform using natural language processing techniques such as text classification, emotion analysis, etc. Specifically, keywords in the barrage or the comment can be extracted, and classification can be performed according to emotion tendencies of the keywords, so that preference degrees of users on different video contents can be known.
Specifically, referring further to fig. 5, a schematic refinement flow chart of the above step S300 is shown. It will be appreciated that the flowchart is not intended to limit the order in which the steps are performed. Some of the steps in the flow chart may be added or subtracted as desired. In this embodiment, the process of determining the user interest classification based on the behavior analysis model in the step S300 specifically includes:
s3000, acquiring a barrage or comment sent by a user when watching a video.
S3002, extracting keywords in the barrage or the comments through text classification.
S3004, carrying out emotion tendency analysis according to the keywords, and determining the preference degree of the user for different video contents.
S3006, determining user interest classification according to commodity information corresponding to the video content.
Through the emotion tendency analysis, the preference degree of the user on the video or a certain fragment thereof can be judged, and then commodity information corresponding to the video or the video fragment is acquired, so that commodities interested by the user can be known, and the interest classification of the user can be determined.
In another alternative embodiment, the viewing behavior analysis model may further assign different weights to different interactions, and determine the user interest classification according to the interactions and their corresponding weights generated when the user views the grass video, the video with goods, and the evaluation video on the current platform.
Specifically, referring further to fig. 6, another refinement flow chart of the above step S300 is shown. It will be appreciated that the flowchart is not intended to limit the order in which the steps are performed. Some of the steps in the flow chart may be added or subtracted as desired. In this embodiment, the determining the user interest classification based on the behavior analysis model in the step S300 may further include:
s3001, obtaining interaction behavior when a user watches video. The interaction behavior comprises the actions of bullet screen sending, comment, praying, coin-in, forwarding, collection, sharing and the like.
S3003, determining the preference degree of the user for different video contents according to the weights of the various interaction behaviors.
S3005, determining user interest classification according to commodity information corresponding to the video content.
The score of the preference degree of the user on the video content can be obtained through the weighted calculation of various interaction behaviors generated when the user watches the video or a certain segment of the video, and then commodity information corresponding to the video or the video segment is obtained, so that the commodities interested by the user can be known, and the interest classification of the user can be determined.
Through the various modes, including comprehensive consideration of indexes such as browsing times, jump records, interaction behaviors and the like, the interest degree score of the user on a certain commodity classification can be calculated, so that the interest classification of the user is finally determined.
Returning to fig. 4, S302 analyzes the commodity type corresponding to each live broadcast according to the live broadcast with the forecast of each live broadcast.
In this embodiment, text analysis and content understanding can be performed on live broadcast and live broadcast previews of each site broadcast by a live broadcast owner through a natural language processing technology, and a commodity type corresponding to each site of live broadcast is determined.
S304, matching the user interest classification with the commodity type of each live broadcast.
And judging the correlation between the commodity type of each live broadcast and the interest classification of the user for the live broadcast with the cargo scene to be broadcast. Based on the commodity classification matching model, text mining techniques, such as word bag models, TF-IDF, etc., may be used to match the user interest classification with the commodity type of each live broadcast. Specifically, the description text of the commodity may be represented in a vectorization manner, and then a similarity (matching degree) between the user interest classification and the commodity type is calculated, so that the commodity type related to the user interest is found.
In addition, collaborative filtering algorithms, such as collaborative filtering based on users, collaborative filtering based on content, etc., may also be used to match the types of goods that may be of interest to the user based on the user's interest classification and interactive behavior analysis results. These algorithms are able to provide personalized merchandise recommendations by analyzing the user's historical behavior and similarity to other users.
S306, recommending live broadcast with-cargo place times to the user according to the matching result.
And sequencing the forecast live broadcast with the cargo field according to the matching degree obtained in the last step. Therefore, the N live broadcast with the top-ranked orders can be recommended to the user, wherein N is a preset positive integer. In other embodiments, a matching degree threshold may be set, and then the live-broadcast on-demand time with the calculated matching degree reaching the threshold may be recommended to the user.
In an alternative embodiment, in addition to considering the user interest classification, other factors such as the hotness of the live room, whether the live time period user is idle, etc. need to be considered in recommending to the user. Specifically, firstly, the playing information of each live broadcast is obtained, wherein the playing information comprises the hotness of a live broadcast room, the live broadcast time period and the like. And then, recommending live broadcast with the goods yard to the user by combining the matching result and the playing information. For example, live orders with high hotness in the live rooms are recommended preferentially in addition to the ranking according to the matching degree. For another example, according to the historical behavior data of the user, judging an idle time period of the live broadcast watched by the user frequently, and recommending the live broadcast time which is matched with the idle time period with priority. In practical application, corresponding weights and formulas can be formulated according to service requirements.
In a word, through the application of big data technology, the interest preference of the user can be deeply known, personalized recommendation service is provided, and accurate live broadcast and live delivery pushing and reservation functions are realized through technical means.
According to the live broadcast recommendation method, according to the commodity browsing record of the user in the live broadcast room and the record of jumping to the third-party electronic commerce platform, and the actions of sending a barrage, praying, inserting coins, forwarding, collecting, sharing and one-key three-link generated when the user watches the grass-carrying video, the live broadcast video and the evaluation video, the user can deduce which commodity classification is interested in recently, and then the classification is matched with a plurality of live broadcast with goods occasions to be released, so that reservation is recommended to the user, the accuracy and pertinence of live broadcast with goods recommendation are improved, and the user experience is improved.
Fig. 7 is a schematic flow chart of an embodiment of the present application. In order to explain the above steps of the method in more detail, specific embodiments are described below as examples. It will be appreciated by those skilled in the art that the following detailed description is not intended to limit the inventive concepts of the present invention, and that suitable content distribution and expansion may be readily implemented by those skilled in the art in light of the detailed description of the embodiments described below.
(1) In a preset time period, when a user watches live broadcast with goods, actions such as opening a small Huang Che list (a live broadcast room commodity list), opening a half-screen page (a commodity detail page), jumping from the half-screen page to a third-party e-commerce platform such as Taobao, jindong, spell and the like may be generated. In addition, the user may watch grass planting video, cargo carrying video, evaluation video, and act of sending barrages, praise, coin-in, forwarding, collecting, sharing, one-key three-connection, etc. on the current platform in the preset time period.
(2) The background server side obtains the behaviors through big data, and judges that a user is possibly interested in certain commodity types based on a recommendation algorithm.
(3) The background server transmits the user interest classification result to the live broadcast with goods server, the live broadcast with goods server searches out live broadcast with goods occasions containing the user interest classification, and when the user opens the live broadcast room next time, or the live broadcast with goods occasions are displayed to the user in a pushing mode, and a reservation function is provided.
Example III
As shown in fig. 8, a hardware architecture diagram of an electronic device 20 according to a third embodiment of the present application is provided. In this embodiment, the electronic device 20 may include, but is not limited to, a memory 21, a processor 22, and a network interface 23, which may be communicatively connected to each other through a system bus. It should be noted that fig. 8 only shows an electronic device 20 having components 21-23, but it should be understood that not all of the illustrated components are required to be implemented and that more or fewer components may alternatively be implemented. In this embodiment, the electronic device 20 may be the server.
The memory 21 includes at least one type of readable storage medium including flash memory, hard disk, multimedia card, card memory (e.g., SD or DX memory, etc.), random Access Memory (RAM), static Random Access Memory (SRAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), programmable Read Only Memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the storage 21 may be an internal storage unit of the electronic device 20, such as a hard disk or a memory of the electronic device 20. In other embodiments, the memory 21 may also be an external storage device of the electronic apparatus 20, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the electronic apparatus 20. Of course, the memory 21 may also include both an internal memory unit and an external memory device of the electronic apparatus 20. In this embodiment, the memory 21 is generally used for storing an operating system and various application software installed on the electronic device 20, such as program codes of the live broadcast recommendation system 60. Further, the memory 21 may be used to temporarily store various types of data that have been output or are to be output.
The processor 22 may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 22 is generally used to control the overall operation of the electronic device 20. In this embodiment, the processor 22 is configured to execute the program code or process data stored in the memory 21, for example, to execute the live recommendation system 60.
The network interface 23 may comprise a wireless network interface or a wired network interface, which network interface 23 is typically used for establishing a communication connection between the electronic apparatus 20 and other electronic devices.
Example IV
As shown in fig. 9, a block diagram of a live recommendation system 60 according to a fourth embodiment of the present application is provided. The live recommendation system 60 may be partitioned into one or more program modules that are stored in a storage medium and executed by one or more processors to perform the embodiments of the present application. Program modules in the embodiments of the present application refer to a series of computer program instruction segments capable of implementing specific functions, and the following description specifically describes the functions of each program module in the embodiments.
In this embodiment, the live recommendation system 60 includes:
the determining module 600 is configured to determine, based on the interest classification model, a user interest classification according to the commodity browsing record of the user on the current platform and the record of jumping to the third party e-commerce platform.
In this embodiment, a commodity browsing record of the live broadcasting room with the current platform and a record of jumping to a third party e-commerce platform in a preset period (for example, one week) of the user are obtained. For example, the user views live broadcast with goods in a live broadcast room of the current platform, clicks on goods in the goods list, opens a goods detail page, and successfully jumps from the goods detail page to a third party e-commerce platform such as Taobao, jindong, spell-up-multiple. From these behavior records, the interest classification model may be constructed, analyzing the user's interest classification.
Specifically, firstly, commodity characteristic information corresponding to the commodity browsing record and the skip record is taken.
And according to the commodity browsing record and the jump record, corresponding commodity characteristic information can be obtained. For example, aiming at the behavior record of a user opening a certain commodity detail page, commodity information corresponding to the commodity detail page is obtained. The commodity characteristic information includes commodity names, brands, prices, and the like.
The merchandise browsing record and the jump record are then converted into behavior feature vectors.
And converting the behavior of the user into a feature vector according to the commodity browsing record and the jump record. The specific transformation method can adopt the existing eigenvector transformation algorithm, and is not limited herein.
And finally, inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model, and predicting the interest degree of the user on different commodity classifications.
The interest classification model may use machine learning algorithms such as decision trees, random forests, support vector machines, and the like. The interest degree of the user for classifying different commodities can be predicted by training the interest classification model. Inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model can output the prediction result of the user interest classification, namely the commodity types which are interested by the user, and the degree of interest.
The determining module 600 is further configured to determine, based on the behavioral analysis model, a user interest classification according to an interactive behavior of the user watching the video on the current platform.
In an alternative embodiment, the viewing behavior analysis model may analyze the bullet screen or comment generated by the user while viewing the grass video, the cargo video, the review video on the current platform using natural language processing techniques such as text classification, emotion analysis, etc. Specifically, keywords in the barrage or the comment can be extracted, and classification can be performed according to emotion tendencies of the keywords, so that preference degrees of users on different video contents can be known.
In another alternative embodiment, the viewing behavior analysis model may further assign different weights to different interactions, and determine the user interest classification according to the interactions and their corresponding weights generated when the user views the grass video, the video with goods, and the evaluation video on the current platform.
Through the various modes, including comprehensive consideration of indexes such as browsing times, jump records, interaction behaviors and the like, the interest degree score of the user on a certain commodity classification can be calculated, so that the interest classification of the user is finally determined.
And the analysis module 602 is used for analyzing the commodity type corresponding to each live broadcast according to the live broadcast with the goods forecast of each live broadcast.
In this embodiment, text analysis and content understanding can be performed on live broadcast and live broadcast previews of each site broadcast by a live broadcast owner through a natural language processing technology, and a commodity type corresponding to each site of live broadcast is determined.
And the matching module 604 is used for matching the user interest classification with the commodity type of each live broadcast.
And judging the correlation between the commodity type of each live broadcast and the interest classification of the user for the live broadcast with the cargo scene to be broadcast. Based on the commodity classification matching model, text mining techniques, such as word bag models, TF-IDF, etc., may be used to match the user interest classification with the commodity type of each live broadcast. Specifically, the description text of the commodity may be represented in a vectorization manner, and then a similarity (matching degree) between the user interest classification and the commodity type is calculated, so that the commodity type related to the user interest is found.
In addition, collaborative filtering algorithms, such as collaborative filtering based on users, collaborative filtering based on content, etc., may also be used to match the types of goods that may be of interest to the user based on the user's interest classification and interactive behavior analysis results. These algorithms are able to provide personalized merchandise recommendations by analyzing the user's historical behavior and similarity to other users.
And the recommending module 606 is used for recommending live broadcast with-cargo occasions to the user according to the matching result.
And sequencing the forecast live broadcast with the cargo field according to the matching degree obtained in the last step. Therefore, the N live broadcast with the top-ranked orders can be recommended to the user, wherein N is a preset positive integer. In other embodiments, a matching degree threshold may be set, and then the live-broadcast on-demand time with the calculated matching degree reaching the threshold may be recommended to the user.
In an alternative embodiment, in addition to considering the user interest classification, other factors such as the hotness of the live room, whether the live time period user is idle, etc. need to be considered in recommending to the user. Specifically, firstly, the playing information of each live broadcast is obtained, wherein the playing information comprises the hotness of a live broadcast room, the live broadcast time period and the like. And then, recommending live broadcast with the goods yard to the user by combining the matching result and the playing information. For example, live orders with high hotness in the live rooms are recommended preferentially in addition to the ranking according to the matching degree. For another example, according to the historical behavior data of the user, judging an idle time period of the live broadcast watched by the user frequently, and recommending the live broadcast time which is matched with the idle time period with priority. In practical application, corresponding weights and formulas can be formulated according to service requirements.
In a word, through the application of big data technology, the interest preference of the user can be deeply known, personalized recommendation service is provided, and accurate live broadcast and live delivery pushing and reservation functions are realized through technical means.
According to the live broadcast recommendation system provided by the embodiment, according to the commodity browsing record of the user in the live broadcast room and the record of jumping to the third party e-commerce platform, and the actions of sending a barrage, praying, inserting coins, forwarding, collecting, sharing and one-key three-connection generated when the user watches the grass-carrying video, the live broadcast video and the evaluation video, the user can deduce which commodity classification is interested in recently, and then the classification is matched with a plurality of live broadcast with goods occasions to be released, so that reservation is recommended to the user, the accuracy and pertinence of live broadcast with goods recommendation are improved, and the user experience is improved.
Example five
The present application also provides another embodiment, namely, a computer readable storage medium storing a live recommendation program executable by at least one processor to cause the at least one processor to perform the steps of the live recommendation method as described above.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
It will be apparent to those skilled in the art that the modules or steps of the embodiments of the application described above may be implemented in a general purpose computing device, they may be concentrated on a single computing device, or distributed across a network of computing devices, they may alternatively be implemented in program code executable by computing devices, so that they may be stored in a storage device for execution by computing devices, and in some cases, the steps shown or described may be performed in a different order than what is shown or described, or they may be separately fabricated into individual integrated circuit modules, or multiple modules or steps of them may be fabricated into a single integrated circuit module. Thus, embodiments of the present application are not limited to any specific combination of hardware and software.
The foregoing is only the preferred embodiments of the present application, and is not intended to limit the scope of the embodiments of the present application, and all equivalent structures or equivalent processes using the descriptions of the embodiments of the present application and the contents of the drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the embodiments of the present application.
Claims (11)
1. A live recommendation method, the method comprising:
determining interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform;
analyzing the commodity type corresponding to each live broadcast according to live broadcast with goods forecast of each live broadcast;
matching the user interest classification with the commodity type of each live broadcast;
and recommending live broadcast with-cargo site times to the user according to the matching result.
2. The live recommendation method of claim 1, further comprising, prior to matching the user interest categories with the merchandise types of each live:
and determining the user interest classification based on the behavior analysis model according to the interactive behavior of the user watching the video on the current platform.
3. The live recommendation method according to claim 1 or 2, wherein the determining the user interest classification based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third party e-commerce platform comprises:
acquiring commodity characteristic information corresponding to the commodity browsing record and the jump record;
converting the commodity browsing records and the jump records into behavior feature vectors;
and inputting the commodity characteristic information and the behavior characteristic vector into the interest classification model, and predicting the interest degree of the user on different commodity classifications.
4. The live broadcast recommendation method according to claim 1 or 2, wherein the analyzing the commodity type corresponding to each live broadcast according to live broadcast with a forecast of each live broadcast comprises:
and carrying out text analysis and content understanding on live broadcast with goods forenotice of each occasion through a natural language processing technology, and determining the commodity type corresponding to each live broadcast.
5. The live recommendation method according to claim 2, wherein the determining the user interest classification based on the behavior analysis model according to the interactive behavior of the user watching the video on the current platform comprises:
acquiring a barrage or comment sent by a user when watching a video;
extracting keywords in the barrage or the comments through text classification;
carrying out emotion tendency analysis according to the keywords, and determining preference degrees of users on different video contents;
and determining the interest classification of the user according to the commodity information corresponding to the video content.
6. The live recommendation method of claim 5, wherein determining the user interest classification based on the behavior analysis model based on the interactive behavior of the user viewing the video on the current platform further comprises:
the method comprises the steps of obtaining interaction behaviors of a user when watching video, wherein the interaction behaviors comprise bullet screen sending, comment sending, praying, coin inserting, forwarding, collecting and sharing behaviors;
determining the preference degree of the user for different video contents according to the weights of various interaction behaviors;
and determining the interest classification of the user according to the commodity information corresponding to the video content.
7. The live broadcast recommendation method according to claim 1 or 2, wherein the recommending live broadcast on-demand to the user according to the matching result comprises:
sequencing the predicted live broadcast with the goods orders according to the matching degree, and recommending N live broadcast with the goods orders which are ranked forward to the user, wherein N is a preset positive integer.
8. The live recommendation method according to claim 7, wherein the recommending live on-demand place to the user according to the matching result further comprises:
acquiring playing information of each live broadcast, wherein the playing information comprises hotness of a live broadcast room and a live broadcast time period;
and recommending live broadcast with goods yard times to the user by combining the matching result and the playing information.
9. A live recommendation system, the system comprising:
the determining module is used for determining interest classification of the user based on the interest classification model according to the commodity browsing record of the user on the current platform and the record of jumping to the third-party e-commerce platform;
the analysis module is used for analyzing the commodity type corresponding to each live broadcast according to live broadcast with goods forecast of each live broadcast;
the matching module is used for matching the user interest classification with the commodity type of each live broadcast;
and the recommendation module is used for recommending live broadcast on-site times to the user according to the matching result.
10. An electronic device, the electronic device comprising: memory, a processor and a live recommendation program stored on the memory and executable on the processor, which when executed by the processor implements the live recommendation method of any of claims 1 to 8.
11. A computer readable storage medium, wherein a live recommendation program is stored on the computer readable storage medium, which when executed by a processor implements a live recommendation method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311321592.2A CN117407598A (en) | 2023-10-12 | 2023-10-12 | Live broadcast recommendation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311321592.2A CN117407598A (en) | 2023-10-12 | 2023-10-12 | Live broadcast recommendation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117407598A true CN117407598A (en) | 2024-01-16 |
Family
ID=89497240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311321592.2A Pending CN117407598A (en) | 2023-10-12 | 2023-10-12 | Live broadcast recommendation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117407598A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118113942A (en) * | 2024-03-21 | 2024-05-31 | 北京振轩网络信息技术有限公司 | Intelligent content recommendation method and system based on user interaction data |
| CN118227828A (en) * | 2024-04-10 | 2024-06-21 | 北京未来链技术有限公司 | Short video live broadcast private domain e-commerce system based on user interest modeling algorithm |
| CN118411223A (en) * | 2024-03-14 | 2024-07-30 | 北京俱宝盆技术有限公司 | Short video live broadcast private domain e-commerce platform based on data mining algorithm |
-
2023
- 2023-10-12 CN CN202311321592.2A patent/CN117407598A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118411223A (en) * | 2024-03-14 | 2024-07-30 | 北京俱宝盆技术有限公司 | Short video live broadcast private domain e-commerce platform based on data mining algorithm |
| CN118113942A (en) * | 2024-03-21 | 2024-05-31 | 北京振轩网络信息技术有限公司 | Intelligent content recommendation method and system based on user interaction data |
| CN118113942B (en) * | 2024-03-21 | 2024-09-17 | 北京振轩网络信息技术有限公司 | Intelligent content recommendation method and system based on user interaction data |
| CN118227828A (en) * | 2024-04-10 | 2024-06-21 | 北京未来链技术有限公司 | Short video live broadcast private domain e-commerce system based on user interest modeling algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111143610B (en) | Content recommendation method and device, electronic equipment and storage medium | |
| CN110941740B (en) | Video recommendation method and computer-readable storage medium | |
| US10565235B2 (en) | Character based media analytics | |
| US10217058B2 (en) | Predicting interesting things and concepts in content | |
| US8234311B2 (en) | Information processing device, importance calculation method, and program | |
| RU2725659C2 (en) | Method and system for evaluating data on user-element interactions | |
| CN117407598A (en) | Live broadcast recommendation method and system | |
| CN109511015B (en) | Multimedia resource recommendation method, device, storage medium and equipment | |
| US20150186368A1 (en) | Comment-based media classification | |
| KR20160019445A (en) | Incorporating user usage of consumable content into recommendations | |
| CN111552835B (en) | File recommendation method, device and server | |
| US20150120634A1 (en) | Information processing device, information processing method, and program | |
| CN116821475B (en) | Video recommendation method and device based on client data and computer equipment | |
| US11727051B2 (en) | Personalized image recommendations for areas of interest | |
| US8838616B2 (en) | Server device for creating list of general words to be excluded from search result | |
| US20230316106A1 (en) | Method and apparatus for training content recommendation model, device, and storage medium | |
| CN113407772B (en) | Video recommendation model generation method, video recommendation method and device | |
| CN113204699B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN117436956A (en) | Intelligent marketing management system based on advertisement pushing | |
| KR101496181B1 (en) | Methods and apparatuses for a content recommendations using content themes | |
| CN112035740B (en) | Project use time length prediction method, device, equipment and storage medium | |
| Kang et al. | Advertisement Recommendation System Based on User Preference in Online Broadcasting | |
| CN112818231A (en) | Information delivery method and device, electronic equipment and storage medium | |
| CN114861033A (en) | Live broadcast recommendation method, device, equipment and computer readable storage medium | |
| CN117651165B (en) | Video recommendation method and device based on client data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |