CN112733014A - Recommendation method, device, equipment and storage medium - Google Patents
Recommendation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112733014A CN112733014A CN202011612465.4A CN202011612465A CN112733014A CN 112733014 A CN112733014 A CN 112733014A CN 202011612465 A CN202011612465 A CN 202011612465A CN 112733014 A CN112733014 A CN 112733014A
- Authority
- CN
- China
- Prior art keywords
- user
- sample
- characteristic
- model
- historical behavior
- 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 83
- 230000002452 interceptive effect Effects 0.000 claims abstract description 59
- 230000006399 behavior Effects 0.000 claims description 148
- 230000003993 interaction Effects 0.000 claims description 45
- 230000008569 process Effects 0.000 claims description 37
- 238000012549 training Methods 0.000 claims description 36
- 238000004891 communication Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 7
- 230000014759 maintenance of location Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 230000003997 social interaction Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000000505 pernicious effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
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
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/44—Browsing; Visualisation therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
The application relates to a recommendation method, a recommendation device, recommendation equipment and a storage medium, wherein the method comprises the following steps: acquiring historical behavior information of a current user; judging whether the user belongs to an interactive group or not according to the historical behavior information; acquiring historical behavior information of a current user; and judging whether the user belongs to an interactive group or not according to the historical behavior information, and when the user is judged to belong to the interactive group, acquiring a first characteristic of the user, and recommending first multimedia data to the user according to the first characteristic, wherein the first characteristic is the interactive behavior of the user and the multimedia characteristic corresponding to the user. The method and the device are used for recommending the corresponding multimedia data for the users of the interactive group.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a recommendation method, apparatus, device, and storage medium.
Background
With the advent of the information age, mass data contents, such as video contents and text contents, are generated every day, and various businesses pay more and more attention to the preference of users so as to better provide personalized contents for the users. Among other things, the accuracy of personalized recommendations increasingly affects user retention and user stickiness of a product.
In addition, for the user, besides clicking and browsing the recommended content, the user may have many interactive behaviors, such as praise, share, and collect. However, many users may stop clicking or watching due to characters, habits and the like, and only some users may have further interactive behaviors such as praise, comment and the like.
How to perform personalized recommendation aiming at users of interactive groups is an urgent problem to be solved.
Disclosure of Invention
The application provides a recommendation method, a recommendation device, recommendation equipment and a storage medium, which are used for recommending corresponding multimedia data for users of an interactive group.
In a first aspect, the present application provides a recommendation method, including:
acquiring historical behavior information of a current user;
judging whether the user belongs to an interactive group or not according to the historical behavior information;
when the user is judged to belong to the interaction group, acquiring a first characteristic of the user, and recommending first multimedia data to the user according to the first characteristic;
the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user.
Optionally, the method further comprises:
when the user is judged not to belong to the interaction group, acquiring a second characteristic of the user, and recommending second multimedia data to the user according to the second characteristic;
the second characteristic is a user portrait of the user, a user behavior and a multimedia characteristic corresponding to the user.
Optionally, judging whether the user belongs to an interactive group according to the historical behavior information includes:
inputting the historical behavior information into a discrimination model, extracting the interaction behavior of the user corresponding to the historical behavior information through the discrimination model, judging whether the user belongs to the interaction group according to the interaction behavior, and outputting a judgment result.
Optionally, judging whether the user belongs to an interactive group according to the historical behavior information includes:
inputting the historical behavior information into a pre-trained discrimination model to obtain the judgment result;
wherein, the training process of the discriminant model comprises the following steps:
obtaining a historical behavior information sample set of a first sample user, wherein the historical behavior information sample set comprises: the historical behavior sample information of N first sample users and a preset user mark of each first sample user are used for indicating whether the first sample users belong to the interaction group, and N is an integer greater than or equal to 1;
respectively executing the following training processes on each piece of historical behavior sample information in the historical behavior information sample set: inputting the historical behavior sample information into an initial discrimination model, extracting an interaction behavior sample corresponding to the historical behavior sample information through the initial discrimination model, judging whether the first sample user belongs to an interaction group according to the interaction behavior sample, and outputting a model judgment result;
calculating a first consistency rate of the model judgment result output by the initial judgment model and a preset user mark of the first sample user;
and if the first consistency rate is not greater than a first preset threshold, after adjusting parameters in the initial discrimination model, repeating the training process until the first consistency rate is greater than the first preset threshold, and taking the initial discrimination model as the final discrimination model.
Optionally, obtaining a first feature of the user, and recommending, according to the first feature, first multimedia data to the user includes:
inputting the first feature to a multi-object model;
calculating the weight of each parameter in the first characteristic through the multi-target model, and outputting the score corresponding to each parameter in the first characteristic according to the weight, wherein the weight and the score are in a direct proportion relation;
sorting the scores corresponding to the parameters respectively, and recommending the first multimedia data to the user according to a sorting result;
the score is used for representing the interest degree of the user corresponding to each parameter in the first characteristic.
Optionally, the training process of the multi-objective model includes:
obtaining a first feature sample set of a second sample user belonging to the interaction group, wherein the first feature sample set comprises: m first characteristic samples of the second sample users and preset scores corresponding to all parameters in each first characteristic sample respectively, wherein M is an integer greater than or equal to 1;
respectively executing the following training process on each first feature sample in the first feature sample set: inputting the first characteristic sample into an initial multi-target model, calculating the weight of each parameter in the first characteristic sample through the initial multi-target model, and outputting the score corresponding to each parameter in the first characteristic sample according to the weight;
calculating the scores respectively corresponding to the parameters output by the initial multi-target model and a second consistent rate of the preset scores;
and if the second consistency rate is not greater than a second preset threshold, after adjusting parameters in the initial multi-target model, repeating the training process until the second consistency rate is greater than the second preset threshold, and taking the initial multi-target model as the final multi-target model.
Optionally, obtaining a second feature of the user, and recommending, according to the second feature, second multimedia data to the user, includes:
inputting the second characteristics into a reference model, analyzing the second characteristics through the reference model, and outputting the number and the playing time length of multimedia data;
recommending the second multimedia data to the user according to the number of the multimedia data and the playing time length;
wherein the training process of the reference model comprises:
obtaining a second feature sample set of a third sample user not belonging to the interaction group, wherein the second feature sample set comprises: second feature samples of Q third sample users, and a preset number and a preset playing time of multimedia data corresponding to each second feature sample, wherein Q is an integer greater than or equal to 1;
respectively executing the following training process on each second feature sample in the second feature sample set: inputting the second characteristic sample into an initial reference model, analyzing the second characteristic sample through the initial reference model, and outputting the number and the playing time length of the multimedia data corresponding to the second characteristic sample;
calculating a third coincidence rate of the number output by the initial reference model and the preset number and a fourth coincidence rate of the playing time length output by the initial reference model and the preset playing time length;
and if the third consistency rate and the fourth consistency rate are not greater than a third preset threshold, after adjusting parameters in the initial reference model, repeating the training process until the third consistency rate and the fourth consistency rate are greater than the third preset threshold, and taking the initial reference model as the final reference model.
In a second aspect, the present application provides a recommendation apparatus, comprising:
the acquisition module is used for acquiring historical behavior information of a current user;
the judging module is used for judging whether the user belongs to an interactive group or not according to the historical behavior information;
the recommending module is used for acquiring a first characteristic of the user when the user is judged to belong to the interactive group, and recommending first multimedia data to the user according to the first characteristic;
the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user.
In a third aspect, the present application provides an electronic device, comprising: the system comprises a processor, a communication component, a memory and a communication bus, wherein the processor, the communication component and the memory are communicated with each other through the communication bus; the memory for storing a computer program; the processor is configured to execute the program stored in the memory to implement the recommendation method of the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the recommendation method of the first aspect.
Compared with the prior art, the technical scheme provided by the embodiment of the application has the following advantages: the method provided by the embodiment of the application acquires the historical behavior information of the current user, judges whether the user belongs to an interactive group or not according to the historical behavior information, when the user is judged to belong to the interactive group, the first characteristic of the user is obtained, and according to the first characteristic, recommending first multimedia data to a user, wherein the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user, judging whether the current user belongs to an interaction group, and further acquiring the characteristic corresponding to the user when judging that the current user belongs to the interaction group, the recommendation of the multimedia data is carried out according to the acquired characteristics, the recommendation of the corresponding multimedia data is realized for the users of the interactive group, and, when the interactive behavior of the user is improved, the exposure rate of the multimedia data is improved due to the high interactive behavior of the user, and the retention rate of the user is also improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a recommendation method in an embodiment of the present application;
FIG. 2 is a schematic diagram of a training process of a discriminant model in an embodiment of the present application;
FIG. 3 is a diagram illustrating a training process of a multi-objective model according to an embodiment of the present application;
FIG. 4 is a diagram illustrating a training process of a reference model in an embodiment of the present application;
FIG. 5 is a flowchart illustrating a recommendation method according to an embodiment of the present application;
FIG. 6 is a schematic structural diagram of a recommendation device in an embodiment of the present application;
fig. 7 is a schematic structural diagram of an electronic device in an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The embodiment of the present application provides a recommendation method, which may be applied to a terminal, for example, a mobile phone, a computer, a tablet, a television, and the like, or may be applied to an application installed in the terminal, for example, a video application, a news application, and the like, or may be applied to a server, and the method is specifically implemented as shown in fig. 1:
the method is described below by taking the application of the method to video applications as an example, but of course, the method is only illustrated here and is not intended to limit the scope of the present application, and the method is not listed here. Moreover, other examples in the present application are not intended to limit the scope of the present application, and thus are not described in detail.
Specifically, when a user opens a video application program, the application program obtains historical behavior information of the user according to a unique identifier of the current user, where the unique identifier may be an account number of the user logging in the application program. The method comprises the following steps of describing historical behavior information of a user from the perspective of a video publisher and a video viewer, wherein the historical behavior information comprises the following steps: historical behavior information of video publishers and historical behavior information of video viewers. Of course, the two users do not conflict with each other in terms of angle, and the video publisher may also be a video viewer, and the video viewer may also be a video publisher, that is, one user may make and publish a video and may also watch a video.
The historical behavior information of the video publisher comprises: a video publishing behavior.
The historical behavior information of the video viewer includes: basic behavior and interactive behavior. Wherein the basic behaviors include: viewing behavior such as video clicks and video swipes, etc. Additionally, video viewing information is also included in the viewing behavior, the video viewing information including: viewing duration, viewing frequency, number of videos viewed, viewing period, etc.
The interactive behaviors include: the method comprises the following steps of social interaction behaviors such as bullet screen launching behaviors, comment behaviors, sharing behaviors and praise behaviors, preference interaction behaviors such as downloading behaviors and collecting behaviors.
In addition, it is also necessary to obtain the user profile of the user, such as the age, sex, city, academic calendar, and attributes of the preference video, while obtaining the historical behavior information of the user, wherein the attributes of the preference video include: the types of the videos include a fun video, a face video, a street video, a pernicious video, an animation video, a cate video, a favorite video and the like, and the duration of the videos.
And 102, judging whether the user belongs to an interactive group or not according to the historical behavior information.
Specifically, the interaction group is a high-activity group, and users belonging to the group can actively participate in interaction while watching videos.
In one embodiment, historical behavior information is input to a discriminant model; and extracting the interaction behavior of the user corresponding to the historical behavior information through the discrimination model, judging whether the user belongs to the interaction group according to the extracted interaction behavior, and outputting a judgment result. The discriminant model is obtained by training historical behavior information sample data. For example, when determining that the interactive behaviors such as comment behavior, like behavior, collection behavior, barrage behavior, download behavior, sharing behavior and the like in the historical behavior information of the user respectively reach respective preset times, the discrimination model determines that the user belongs to the interactive group, otherwise, determines that the user does not belong to the interactive group.
For example, the number of times that the user meets the requirement of making comments reaches a first preset number of times, the number of times of praise reaches a second preset number of times, the number of times of collecting videos reaches a third preset number of times, the number of times of publishing a barrage when watching videos reaches a fourth preset number of times, the number of times of downloading videos reaches a sixth preset number of times, if not, it is determined that the user belongs to an interactive group, and if not, it is determined that the user does not belong to the interactive group.
The method and the device adopt the discrimination model to judge whether the user belongs to the interactive group or not, aim to group-divide the user and recommend different videos aiming at the users of different groups so as to meet the watching requirements of the users of different groups and improve the watching duration of the videos.
In a specific embodiment, the historical behavior information is input to a pre-trained discriminant model, and a discriminant result is output, wherein a training process of the discriminant model is specifically shown in fig. 2:
And 204, if the first consistency rate is not greater than the first preset threshold, after adjusting parameters in the initial discrimination model, repeating the training process until the first consistency rate is greater than the first preset threshold, and taking the initial discrimination model as a final discrimination model.
In addition, after the discriminant model is applied, historical behavior information of a first preset time period is collected at regular time, the historical behavior information of the first preset time period is formed into a new historical behavior information sample set of the first sample user, the training process of the discriminant model is repeatedly executed on the new historical behavior information sample set of the first sample user, and the discriminant model is updated successfully until the first consistency rate is greater than a first preset threshold value.
According to the method and the device, the judgment result of the judgment model is more and more accurate through continuously updating the judgment model, the user group division precision is higher and higher, and the retention rate of the user is improved.
In addition, besides the mode of inputting the historical behavior information corresponding to the user into the discrimination model to obtain whether the user belongs to the interactive group, the following mode can be adopted to realize the following steps:
and extracting any one or more preset interactive behaviors from the historical behavior information, respectively counting the occurrence frequency of each extracted interactive behavior, judging that the user belongs to an interactive group when the frequency of each interactive behavior is respectively greater than the preset frequency, and otherwise, judging that the user does not belong to the interactive group. And if any item in the preset interaction behaviors cannot be extracted, judging that the user does not belong to the interaction group.
Specifically, the interactive behavior includes: social interaction behaviors such as bullet screen launching behavior, comment behavior, sharing behavior and like behavior, preference interaction behaviors such as downloading behavior and collection behavior, and like behavior; the multimedia features include: the type of video, the duration of the video, etc.
In one embodiment, a first feature of a user is input to a multi-goal model; calculating the weight of each parameter in the first characteristic through a multi-target model, and outputting the score corresponding to each parameter in the first characteristic according to the weight, wherein the weight and the score are in a direct proportion relation; sorting the scores corresponding to the parameters respectively, and recommending the first multimedia data to the user according to a sorting result; the scores are used for representing the interest degrees of the users corresponding to the parameters in the first characteristics respectively, and the multi-target model is obtained through training of first characteristic sample data of the users belonging to the interaction group. The Multi-objective model may be a Multi-Task model (Modeling Task Relationships in Multi-Task Learning with Multi-gate understanding-of-Experts, MMOE model for short).
For example, the first characteristic includes parameters such as a pop-up behavior, a comment behavior, a sharing behavior, a like behavior, a download behavior, and a collection behavior, and the weight of each parameter is obtained through multi-objective model calculation, for example, the comment behavior accounts for 50%, the like behavior accounts for 30%, the pop-up behavior accounts for 18%, the download behavior accounts for 2%, the sharing behavior accounts for 0%, and the collection behavior accounts for 0%. And further determining the scores of the parameters according to the obtained weights, obviously determining that the scores of the comments and praise behaviors of the user are higher, and determining that the user prefers to issue the comments and praise, so that the videos needing to be popularized and increasing the interactive exposure rate can be recommended to the user on the basis that the recommended videos meet the preference requirements of the user.
In one embodiment, the training process of the multi-objective model is specifically shown in fig. 3:
And 303, calculating the scores respectively corresponding to the parameters output by the initial multi-target model and a second consistent rate of the preset scores.
And 304, if the second consistency rate is not greater than the second preset threshold, after adjusting parameters in the initial multi-target model, repeating the training process until the second consistency rate is greater than the second preset threshold, and taking the initial multi-target model as a final multi-target model.
In addition, after the multi-target model application is regularly acquired and updated, a new first characteristic sample set of a second sample user belonging to the interactive group in a second preset time period is acquired, the training process of the multi-target model is repeatedly executed on the new first characteristic sample set of the second sample user, and the multi-target model is successfully updated until the second consistency rate is larger than a second preset threshold value.
According to the method and the device, the multi-target model is continuously updated, so that the weight of each parameter output by the multi-target model is more and more accurate, and the retention rate of the user is improved.
In addition, besides that the first feature corresponding to the user is input into the multi-objective model to obtain the scores of the parameters in the first feature, and the first multimedia data is recommended to the user according to the obtained scores, the following method can be adopted to realize the following steps:
and calculating the weight of each parameter of the first characteristic, sorting the parameters of the first characteristic from high to low according to the calculated weight to obtain a sorting result, and recommending the first multimedia data to the user according to the sorting result.
In a specific embodiment, if the determination result is negative, the second feature of the user is obtained, and the second multimedia data is recommended to the user according to the second feature, wherein the second feature is a user portrait, a user behavior and a multimedia feature of the user.
Specifically, the user behavior here is the basic behavior of the user.
In a specific embodiment, the second characteristic of the user is input into the reference model, the second characteristic is analyzed through the reference model, and the number and the playing time length of the multimedia data are output; and recommending the second multimedia data to the user according to the number and the playing time of the multimedia data. The number of the multimedia data is the number of videos watched by the user, and the playing time length is the time length corresponding to each video watched by the user and the total time length of the videos watched by the user. The benchmark model is obtained through training of second feature sample data of users not belonging to the interactive group.
In an embodiment, the training process of the reference model is specifically as shown in fig. 4:
In step 403, a third coincidence rate of the number of the initial reference model outputs and the preset number and a fourth coincidence rate of the playing duration of the initial reference model outputs and the preset playing duration are calculated.
And step 404, if the third consistency rate and the fourth consistency rate are not greater than the third preset threshold, after adjusting parameters in the initial reference model, repeating the training process until the third consistency rate and the fourth consistency rate are greater than the third preset threshold, and taking the initial reference model as a final reference model.
In addition, after the reference model is regularly acquired, a new second feature sample set of a third sample user, which does not belong to the interactive group, in a third preset time period is acquired, and the training process of the reference model is repeatedly executed on the new second feature sample set of the third sample user until the third consistency rate and the fourth consistency rate are both greater than a third preset threshold value, and the reference model is successfully updated.
According to the method and the device, the reference model is continuously updated, so that the number of the videos output by the reference model and the playing time of the videos are more and more accurate, the playing time of the videos is prolonged, and the retention rate of users is also improved.
In addition, besides the second feature corresponding to the user is input to the reference model to obtain the number and the playing time length of the multimedia data corresponding to the user, the following method can be adopted to implement:
and according to the second characteristic, counting the multimedia data and the playing time length corresponding to the user, and recommending second multimedia data to the user according to the number of the multimedia data and the playing time length.
The following describes the recommendation method specifically with reference to fig. 5:
And 505, acquiring a second characteristic of the user, and recommending second multimedia data to the user according to the second characteristic through the reference model.
The method provided by the embodiment of the application acquires the historical behavior information of the current user, judges whether the user belongs to an interactive group or not according to the historical behavior information, when the user is judged to belong to the interactive group, the first characteristic of the user is obtained, and according to the first characteristic, recommending first multimedia data to a user, wherein the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user, judging whether the current user belongs to an interaction group, and further acquiring the characteristic corresponding to the user when judging that the current user belongs to the interaction group, the recommendation of the multimedia data is carried out according to the acquired characteristics, the recommendation of the corresponding multimedia data is realized for the users of the interactive group, and, when the interactive behavior of the user is improved, the exposure rate of the multimedia data is improved due to the high interactive behavior of the user, and the retention rate of the user is also improved.
In another embodiment of the present application, a recommendation apparatus is provided, and specific implementation of the apparatus may refer to the description of the method embodiment, and repeated details are not repeated, as shown in fig. 6, the apparatus mainly includes:
the obtaining module 601 is configured to obtain historical behavior information of a current user.
The determining module 602 is configured to determine whether the user belongs to an interactive group according to the historical behavior information.
And the recommending module 603 is configured to, when it is determined that the user belongs to the interactive group, obtain a first feature of the user, and recommend the first multimedia data to the user according to the first feature.
The first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user.
Based on the same concept, an embodiment of the present application further provides an electronic device, as shown in fig. 7, the electronic device mainly includes: a processor 701, a communication component 702, a memory 703 and a communication bus 704, wherein the processor 701, the communication component 702 and the memory 703 communicate with each other via the communication bus 704. The memory 703 stores a program executable by the processor 701, and the processor 701 executes the program stored in the memory 703 to implement the following steps: acquiring historical behavior information of a current user; and judging whether the user belongs to an interactive group or not according to the historical behavior information, and when the user is judged to belong to the interactive group, acquiring a first characteristic of the user, and recommending first multimedia data to the user according to the first characteristic, wherein the first characteristic is the interactive behavior of the user and the multimedia characteristic corresponding to the user.
The communication bus 704 mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus 704 may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 7, but this is not intended to represent only one bus or type of bus.
The communication component 702 is used for communication between the electronic device and other devices described above.
The Memory 703 may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Alternatively, the memory may be at least one memory device located remotely from the processor 701.
The Processor 701 may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like, or may be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components.
In yet another embodiment of the present application, there is also provided a computer-readable storage medium having stored therein a computer program which, when run on a computer, causes the computer to execute the recommendation method described in the above embodiment.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wirelessly (e.g., infrared, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The available media may be magnetic media (e.g., floppy disks, hard disks, tapes, etc.), optical media (e.g., DVDs), or semiconductor media (e.g., solid state drives), among others.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, 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 an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A recommendation method, characterized in that the method comprises:
acquiring historical behavior information of a current user;
judging whether the user belongs to an interactive group or not according to the historical behavior information;
when the user is judged to belong to the interaction group, acquiring a first characteristic of the user, and recommending first multimedia data to the user according to the first characteristic;
the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user.
2. The recommendation method according to claim 1, further comprising:
when the user is judged not to belong to the interaction group, acquiring a second characteristic of the user, and recommending second multimedia data to the user according to the second characteristic;
the second characteristic is a user portrait of the user, a user behavior and a multimedia characteristic corresponding to the user.
3. The recommendation method according to claim 1 or 2, wherein determining whether the user belongs to an interactive group according to the historical behavior information comprises:
inputting the historical behavior information into a discrimination model, extracting the interaction behavior of the user corresponding to the historical behavior information through the discrimination model, judging whether the user belongs to the interaction group according to the interaction behavior, and outputting a judgment result.
4. The recommendation method according to claim 1 or 2, wherein determining whether the user belongs to an interactive group according to the historical behavior information comprises:
inputting the historical behavior information into a pre-trained discrimination model to obtain the judgment result;
wherein, the training process of the discriminant model comprises the following steps:
obtaining a historical behavior information sample set of a first sample user, wherein the historical behavior information sample set comprises: the historical behavior sample information of N first sample users and a preset user mark of each first sample user are used for indicating whether the first sample users belong to the interaction group, and N is an integer greater than or equal to 1;
respectively executing the following training processes on each piece of historical behavior sample information in the historical behavior information sample set: inputting the historical behavior sample information into an initial discrimination model, extracting an interaction behavior sample corresponding to the historical behavior sample information through the initial discrimination model, judging whether the first sample user belongs to an interaction group according to the interaction behavior sample, and outputting a model judgment result;
calculating a first consistency rate of the model judgment result output by the initial judgment model and a preset user mark of the first sample user;
and if the first consistency rate is not greater than a first preset threshold, after adjusting parameters in the initial discrimination model, repeating the training process until the first consistency rate is greater than the first preset threshold, and taking the initial discrimination model as the final discrimination model.
5. The recommendation method according to claim 1, wherein obtaining a first feature of the user, and recommending first multimedia data to the user according to the first feature comprises:
inputting the first feature to a multi-object model;
calculating the weight of each parameter in the first characteristic through the multi-target model, and outputting the score corresponding to each parameter in the first characteristic according to the weight, wherein the weight and the score are in a direct proportion relation;
sorting the scores corresponding to the parameters respectively, and recommending the first multimedia data to the user according to a sorting result;
the score is used for representing the interest degree of the user corresponding to each parameter in the first characteristic.
6. The recommendation method according to claim 5, wherein the training process of the multi-objective model comprises:
obtaining a first feature sample set of a second sample user belonging to the interaction group, wherein the first feature sample set comprises: m first characteristic samples of the second sample users and preset scores corresponding to all parameters in each first characteristic sample respectively, wherein M is an integer greater than or equal to 1;
respectively executing the following training process on each first feature sample in the first feature sample set: inputting the first characteristic sample into an initial multi-target model, calculating the weight of each parameter in the first characteristic sample through the initial multi-target model, and outputting the score corresponding to each parameter in the first characteristic sample according to the weight;
calculating the scores respectively corresponding to the parameters output by the initial multi-target model and a second consistent rate of the preset scores;
and if the second consistency rate is not greater than a second preset threshold, after adjusting parameters in the initial multi-target model, repeating the training process until the second consistency rate is greater than the second preset threshold, and taking the initial multi-target model as the final multi-target model.
7. The recommendation method according to claim 2, wherein obtaining a second feature of the user, and recommending second multimedia data to the user according to the second feature comprises:
inputting the second characteristics into a reference model, analyzing the second characteristics through the reference model, and outputting the number and the playing time length of multimedia data;
recommending the second multimedia data to the user according to the number of the multimedia data and the playing time length;
wherein the training process of the reference model comprises:
obtaining a second feature sample set of a third sample user not belonging to the interaction group, wherein the second feature sample set comprises: second feature samples of Q third sample users, and a preset number and a preset playing time of multimedia data corresponding to each second feature sample, wherein Q is an integer greater than or equal to 1;
respectively executing the following training process on each second feature sample in the second feature sample set: inputting the second characteristic sample into an initial reference model, analyzing the second characteristic sample through the initial reference model, and outputting the number and the playing time length of the multimedia data corresponding to the second characteristic sample;
calculating a third coincidence rate of the number output by the initial reference model and the preset number and a fourth coincidence rate of the playing time length output by the initial reference model and the preset playing time length;
and if the third consistency rate and the fourth consistency rate are not greater than a third preset threshold, after adjusting parameters in the initial reference model, repeating the training process until the third consistency rate and the fourth consistency rate are greater than the third preset threshold, and taking the initial reference model as the final reference model.
8. A recommendation device, comprising:
the acquisition module is used for acquiring historical behavior information of a current user;
the judging module is used for judging whether the user belongs to an interactive group or not according to the historical behavior information;
the recommending module is used for acquiring a first characteristic of the user when the user is judged to belong to the interactive group, and recommending first multimedia data to the user according to the first characteristic;
the first characteristic is the interaction behavior of the user and the multimedia characteristic corresponding to the user.
9. An electronic device, comprising: the system comprises a processor, a communication component, a memory and a communication bus, wherein the processor, the communication component and the memory are communicated with each other through the communication bus;
the memory for storing a computer program;
the processor, executing the program stored in the memory, implementing the recommended method of any one of claims 1-7.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011612465.4A CN112733014A (en) | 2020-12-30 | 2020-12-30 | Recommendation method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011612465.4A CN112733014A (en) | 2020-12-30 | 2020-12-30 | Recommendation method, device, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112733014A true CN112733014A (en) | 2021-04-30 |
Family
ID=75611010
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011612465.4A Pending CN112733014A (en) | 2020-12-30 | 2020-12-30 | Recommendation method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112733014A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113626712A (en) * | 2021-08-19 | 2021-11-09 | 云南腾云信息产业有限公司 | Content determination method and device based on user interaction behavior |
CN113724454A (en) * | 2021-08-25 | 2021-11-30 | 上海擎朗智能科技有限公司 | Interaction method of mobile equipment, device and storage medium |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103686396A (en) * | 2013-11-19 | 2014-03-26 | 乐视致新电子科技(天津)有限公司 | Video sharing method and device |
CN106446078A (en) * | 2016-09-08 | 2017-02-22 | 乐视控股(北京)有限公司 | Information recommendation method and recommendation apparatus |
CN106484777A (en) * | 2016-09-12 | 2017-03-08 | 腾讯科技(深圳)有限公司 | A kind of multimedia data processing method and device |
CN107609198A (en) * | 2017-10-20 | 2018-01-19 | 咪咕互动娱乐有限公司 | One kind recommends method, apparatus and computer-readable recording medium |
CN108388624A (en) * | 2018-02-12 | 2018-08-10 | 科大讯飞股份有限公司 | Multimedia messages recommend method and device |
CN110162703A (en) * | 2019-05-13 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Content recommendation method, training method, device, equipment and storage medium |
CN110321455A (en) * | 2019-06-19 | 2019-10-11 | 北京奇艺世纪科技有限公司 | A kind of information-pushing method and device |
CN110781321A (en) * | 2019-08-28 | 2020-02-11 | 腾讯科技(深圳)有限公司 | Multimedia content recommendation method and device |
CN110837598A (en) * | 2019-11-11 | 2020-02-25 | 腾讯科技(深圳)有限公司 | Information recommendation method, device, equipment and storage medium |
CN111010592A (en) * | 2019-12-19 | 2020-04-14 | 上海众源网络有限公司 | Video recommendation method and device, electronic equipment and storage medium |
CN111222030A (en) * | 2018-11-27 | 2020-06-02 | 阿里巴巴集团控股有限公司 | Information recommendation method and device and electronic equipment |
CN111310034A (en) * | 2020-01-23 | 2020-06-19 | 腾讯科技(深圳)有限公司 | Resource recommendation method and related equipment |
CN111723910A (en) * | 2020-06-17 | 2020-09-29 | 腾讯科技(北京)有限公司 | Method and device for constructing multi-task learning model, electronic equipment and storage medium |
-
2020
- 2020-12-30 CN CN202011612465.4A patent/CN112733014A/en active Pending
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103686396A (en) * | 2013-11-19 | 2014-03-26 | 乐视致新电子科技(天津)有限公司 | Video sharing method and device |
CN106446078A (en) * | 2016-09-08 | 2017-02-22 | 乐视控股(北京)有限公司 | Information recommendation method and recommendation apparatus |
CN106484777A (en) * | 2016-09-12 | 2017-03-08 | 腾讯科技(深圳)有限公司 | A kind of multimedia data processing method and device |
CN107609198A (en) * | 2017-10-20 | 2018-01-19 | 咪咕互动娱乐有限公司 | One kind recommends method, apparatus and computer-readable recording medium |
CN108388624A (en) * | 2018-02-12 | 2018-08-10 | 科大讯飞股份有限公司 | Multimedia messages recommend method and device |
CN111222030A (en) * | 2018-11-27 | 2020-06-02 | 阿里巴巴集团控股有限公司 | Information recommendation method and device and electronic equipment |
CN110162703A (en) * | 2019-05-13 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Content recommendation method, training method, device, equipment and storage medium |
CN110321455A (en) * | 2019-06-19 | 2019-10-11 | 北京奇艺世纪科技有限公司 | A kind of information-pushing method and device |
CN110781321A (en) * | 2019-08-28 | 2020-02-11 | 腾讯科技(深圳)有限公司 | Multimedia content recommendation method and device |
CN110837598A (en) * | 2019-11-11 | 2020-02-25 | 腾讯科技(深圳)有限公司 | Information recommendation method, device, equipment and storage medium |
CN111010592A (en) * | 2019-12-19 | 2020-04-14 | 上海众源网络有限公司 | Video recommendation method and device, electronic equipment and storage medium |
CN111310034A (en) * | 2020-01-23 | 2020-06-19 | 腾讯科技(深圳)有限公司 | Resource recommendation method and related equipment |
CN111723910A (en) * | 2020-06-17 | 2020-09-29 | 腾讯科技(北京)有限公司 | Method and device for constructing multi-task learning model, electronic equipment and storage medium |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113626712A (en) * | 2021-08-19 | 2021-11-09 | 云南腾云信息产业有限公司 | Content determination method and device based on user interaction behavior |
CN113724454A (en) * | 2021-08-25 | 2021-11-30 | 上海擎朗智能科技有限公司 | Interaction method of mobile equipment, device and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107832437B (en) | Audio/video pushing method, device, equipment and storage medium | |
CN106651542B (en) | Article recommendation method and device | |
CN110149540B (en) | Recommendation processing method and device for multimedia resources, terminal and readable medium | |
CN108540826B (en) | Bullet screen pushing method and device, electronic equipment and storage medium | |
CN111708901B (en) | Multimedia resource recommendation method and device, electronic equipment and storage medium | |
CN111767429B (en) | Video recommendation method and device and electronic equipment | |
CN109753601B (en) | Method and device for determining click rate of recommended information and electronic equipment | |
US20140279121A1 (en) | Customizable and adjustable pricing of games | |
US20140274355A1 (en) | Dynamic recommendation of games | |
CN111107416B (en) | Bullet screen shielding method and device and electronic equipment | |
CN110737859A (en) | UP main matching method and device | |
CN110851712B (en) | Method, device and computer readable medium for recommending book information | |
US20140274354A1 (en) | Intelligent merchandising of games | |
CN112418920A (en) | Training method of information recommendation model, information recommendation method and device | |
CN112733014A (en) | Recommendation method, device, equipment and storage medium | |
WO2023000491A1 (en) | Application recommendation method, apparatus and device, and computer-readable storage medium | |
JP2011227721A (en) | Interest extraction device, interest extraction method, and interest extraction program | |
CN107656918A (en) | Obtain the method and device of targeted customer | |
CN113297486B (en) | Click rate prediction method and related device | |
CN108062415B (en) | User behavior statistical method and device and terminal equipment | |
CN111177564B (en) | Product recommendation method and device | |
US10237226B2 (en) | Detection of manipulation of social media content | |
CN107341172B (en) | Video profit calculation modeling device and method and video recommendation device and method | |
CN105447087A (en) | Video recommendation method and apparatus | |
CN115834959A (en) | Video recommendation information determination method and device, electronic equipment and medium |
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 |