CN113626679A - Multimedia resource recommendation method, device and storage medium - Google Patents
Multimedia resource recommendation method, device and storage medium Download PDFInfo
- Publication number
- CN113626679A CN113626679A CN202010378667.0A CN202010378667A CN113626679A CN 113626679 A CN113626679 A CN 113626679A CN 202010378667 A CN202010378667 A CN 202010378667A CN 113626679 A CN113626679 A CN 113626679A
- Authority
- CN
- China
- Prior art keywords
- prediction parameter
- account
- parameters
- parameter set
- multimedia resources
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The disclosure discloses a multimedia resource recommendation method, a multimedia resource recommendation device and a storage medium, relates to the technical field of mobile internet, and at least solves the problem that the accuracy of recommending multimedia resources is low in the related technology. In the method, a first prediction parameter set and a second prediction parameter set of candidate multimedia resources are obtained by predicting the interactive operation of a target account on the candidate multimedia resources; and determining screening parameters according to the obtained first prediction parameter set and the second prediction parameter set, and recommending multimedia resources to the target account according to the screening parameters. Therefore, a more accurate recommendation result is generated according to the screening parameters obtained by the first prediction parameter set and the second prediction parameter set, and the accuracy of recommending the multimedia resources to the target account is further improved.
Description
Technical Field
The present disclosure relates to the field of mobile internet technologies, and in particular, to a multimedia resource recommendation method, apparatus, and storage medium.
Background
With the development of the times, electronic devices have become a part of life, and various application software is generally downloaded and installed in the electronic devices in order to meet the use requirements of users. The user can watch videos, listen to music, etc. through the application software. In order to recommend multimedia resources such as videos, music or news which are interesting to the user, the application software screens the multimedia resources which are interesting to the user from a resource library through a recommendation algorithm.
However, the existing recommendation algorithm places the focus on the behavior of paying attention, that is, the behavior of paying attention to the publisher of the recommended multimedia resource after the user watches the multimedia resource. Ignoring the resulting attention behavior does not necessarily result in an effective state change, e.g., the user pays attention to the publisher and does not watch other multimedia assets of the publisher. Therefore, the existing recommendation method has no way to distinguish which behaviors belong to the behaviors that effectively cause the state change, thereby reducing the accuracy of recommending the multimedia resources.
Disclosure of Invention
The embodiment of the disclosure provides a multimedia resource recommendation method, a multimedia resource recommendation device and a storage medium, so as to improve the accuracy of recommending multimedia resources to a user.
According to a first aspect of the embodiments of the present disclosure, a multimedia resource recommendation method is provided, including:
acquiring a candidate multimedia resource set of a target account; the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account;
generating a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, wherein the first prediction parameter set is used for recording each prediction parameter of the target account for performing interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account;
obtaining screening parameters of the candidate multimedia resources based on the first prediction parameter set and the second prediction parameter set;
and screening multimedia resources from the candidate multimedia resource set according to the screening parameters and recommending the multimedia resources to the target account.
In a possible implementation manner, the obtaining the screening parameters of the candidate multimedia resources based on the first prediction parameter set and the second prediction parameter set includes:
fusing all the prediction parameters in the second prediction parameter set to obtain correction parameters;
and correcting the prediction parameters in the first prediction parameter set according to the correction parameters to obtain the screening parameters of the candidate multimedia resources.
In a possible implementation manner, the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources includes:
multiplying the prediction parameters associated with the correction parameters in the first prediction parameter set by the correction parameters to obtain corrected prediction parameters; the prediction parameter associated with the correction parameter is a prediction parameter for the target account to perform a specific operation on the publishing account through the candidate multimedia resource, wherein the specific operation is one of the interactive operations;
and carrying out weighted summation calculation on the corrected prediction parameters and other prediction parameters in the first prediction parameter set according to preset coefficients to obtain screening parameters of the candidate multimedia resources.
In a possible implementation manner, the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources includes:
and performing weighted summation calculation on each prediction parameter in the first prediction parameter set and the correction parameter according to a preset coefficient to obtain the screening parameters of the candidate multimedia resources.
In a possible implementation manner, the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources includes:
carrying out feature extraction on the correction parameters to obtain correction features; and;
performing feature extraction on each prediction parameter in the first prediction parameter set to obtain the feature of each prediction parameter;
and performing linear calculation on the characteristics of each prediction parameter and the correction characteristics to obtain the screening parameters of the candidate multimedia resources.
In a possible implementation, the second prediction parameter set corresponding to the candidate multimedia resource is determined by:
performing feature extraction on the account information of the release account to obtain release account features; and;
carrying out feature extraction on account information of an associated account set having an association relation with the release account to obtain associated account features; the associated account set comprises at least two associated accounts, the account information of the associated account set comprises account information of each associated account, and interactive operation executed by each associated account on the multimedia resource of the release account;
determining the second prediction parameter set of the interactive operation performed by the target account on the issuing account according to the issuing account characteristic and the associated account characteristic.
In a possible implementation manner, the filtering, according to the filtering parameter, a multimedia resource from the candidate multimedia resource set to recommend to the target account includes:
determining multimedia resources to be recommended from the candidate multimedia resource set according to the screening parameters;
if the number of the same type of multimedia resources in the determined multimedia resources to be recommended is greater than a preset threshold value, screening the multimedia resources in the same type of which the number is greater than the preset threshold value according to the screening parameters; the types of the multimedia resources are obtained by classifying the contents of the multimedia resources;
and recommending the screened multimedia resources and other multimedia resources to be recommended to the target account.
According to a second aspect of the embodiments of the present disclosure, there is provided a multimedia resource recommendation apparatus, including:
an acquisition unit configured to perform acquisition of a set of candidate multimedia resources of a target account; the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account;
a prediction parameter set determining unit configured to perform generation of a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, where the first prediction parameter set is used to record each prediction parameter of the target account performing an interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account;
a screening parameter determining unit configured to perform screening parameter obtaining on the candidate multimedia resource based on the first prediction parameter set and the second prediction parameter set;
and the recommending unit is configured to perform screening of multimedia resources from the candidate multimedia resource set according to the screening parameters and recommend the multimedia resources to the target account.
In one possible implementation, the determining the screening parameter unit includes:
a fusion subunit, configured to perform fusion processing on each prediction parameter in the second prediction parameter set, so as to obtain a correction parameter;
and the correcting subunit is configured to correct the prediction parameters in the first prediction parameter set according to the correction parameters, so as to obtain the screening parameters of the candidate multimedia resources.
In one possible implementation, the correction subunit includes:
a first correction subunit configured to perform multiplication of a prediction parameter associated with the correction parameter in the first prediction parameter set and the correction parameter to obtain a corrected prediction parameter; the prediction parameter associated with the correction parameter is a prediction parameter for the target account to perform a specific operation on the publishing account through the candidate multimedia resource, wherein the specific operation is one of the interactive operations;
and the screening parameter determining subunit is configured to perform weighted summation calculation on the corrected prediction parameters and other prediction parameters in the first prediction parameter set according to preset coefficients to obtain screening parameters of the candidate multimedia resources.
In one possible implementation, the correction subunit includes:
and the second correction subunit is configured to perform weighted summation calculation on each prediction parameter in the first prediction parameter set and the correction parameter according to a preset coefficient to obtain a screening parameter of the candidate multimedia resource.
In one possible implementation, the correction subunit includes:
a first feature extraction subunit configured to perform feature extraction on the correction parameter to obtain a correction feature;
a second feature extraction subunit, configured to perform feature extraction on each prediction parameter in the first prediction parameter set to obtain a feature of each prediction parameter;
and the third correction subunit is configured to perform linear calculation on the characteristics of the prediction parameters and the correction characteristics to obtain the screening parameters of the candidate multimedia resources.
In a possible implementation manner, the second prediction parameter set corresponding to the candidate multimedia resource is determined by:
the first feature extraction unit is configured to perform feature extraction on the account information of the release account to obtain release account features;
the second feature extraction unit is configured to perform feature extraction on account information of an associated account set having an association relation with the release account to obtain associated account features; the associated account set comprises at least two associated accounts, the account information of the associated account set comprises account information of each associated account, and interactive operation executed by each associated account on the multimedia resource of the release account;
a determine second prediction parameter set unit configured to perform the second prediction parameter set that determines the interaction performed by the target account with the posting account according to the posting account characteristic and the associated account characteristic.
In one possible implementation, the recommending unit includes:
a multimedia resource determining subunit configured to perform determining, according to the filtering parameter, a multimedia resource to be recommended from the candidate multimedia resource set;
the screening subunit is configured to perform screening on the multimedia resources in the same type of which the number is greater than a preset threshold value according to the screening parameters if the number of the multimedia resources in the same type in the determined multimedia resources to be recommended is greater than the preset threshold value; the types of the multimedia resources are obtained by classifying the contents of the multimedia resources;
and the recommending subunit is configured to recommend the filtered multimedia resources and other multimedia resources to be recommended to the target account.
According to a third aspect of the embodiments of the present disclosure, there is provided an electronic apparatus including:
a processor;
a memory for storing the processor-executable instructions;
wherein the processor is configured to execute the instructions to implement a multimedia resource recommendation method;
according to a fourth aspect of the embodiments of the present disclosure, there is provided a storage medium having instructions that, when executed by a processor of an electronic device, enable the electronic device to perform a multimedia resource recommendation method;
according to a fifth aspect of embodiments of the present disclosure, there is provided a computer program product comprising at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the multimedia resource recommendation method provided by the embodiment of the disclosure.
The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects:
predicting the interactive operation of the target account on the candidate multimedia resources to obtain a first prediction parameter set and a second prediction parameter set of the candidate multimedia resources, wherein the second prediction parameter set is used for recording each prediction parameter of the interactive operation of the target account on other multimedia resources published by a publishing account after the social relationship is established between the target account and the publishing account of the candidate multimedia resources; and determining screening parameters according to the obtained first prediction parameter set and the second prediction parameter set, and recommending multimedia resources to the target account according to the screening parameters. Therefore, by acquiring the second prediction parameter set, the prediction parameters of the target account performing interactive operation on other multimedia resources of the release account after paying attention to the release account can be predicted, so that a more accurate recommendation result is generated according to the screening parameters obtained by the first prediction parameter set and the second prediction parameter set, and the accuracy of recommending the multimedia resources to the target account is further improved.
Additional features and advantages of the disclosure will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the disclosure. The objectives and other advantages of the disclosure may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and not to limit the disclosure. In the drawings:
FIG. 1 is a schematic diagram of an application interface in an embodiment of the present disclosure;
FIG. 2 is a flowchart illustrating a multimedia resource recommendation method according to an embodiment of the disclosure;
FIG. 3 is a schematic structural diagram of a multimedia resource recommendation device according to an embodiment of the disclosure;
fig. 4 is a schematic structural diagram of a terminal device in an embodiment of the present disclosure.
Detailed Description
In order to improve the accuracy of recommending multimedia resources to a user, the embodiment of the disclosure provides a multimedia resource recommending method, a multimedia resource recommending device and a storage medium. In order to make the technical solutions of the present disclosure better understood by those of ordinary skill in the art, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.
It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
The technical scheme provided by the embodiment of the disclosure is described below with reference to the accompanying drawings.
In order to improve user stickiness, existing application software generally provides two pages in an application, wherein one page is a recommendation page for recommending multimedia resources to a user, and the other page is a multimedia resource released by an author concerned by the user. Taking the video as an example, the user can view the recommended video on the recommended page, and can also view the video published by the concerned author on the concerned page. As shown in fig. 1, which is a schematic diagram of an application interface. And the user enters the corresponding page by clicking the corresponding touch option. Such as: and if the user clicks the recommendation page, the application displays the recommendation page, and the user clicks the video in the recommendation page to play the recommendation video.
And the application hopes that the user can generate the behavior of paying attention to the video author after browsing the favorite video on the recommended page, and further convert the behavior to the attention page to browse the video, so that the viscosity of the user can be improved, the production enthusiasm of the author can be improved, and the whole content ecology forms a good cycle.
However, in the existing recommendation scheme, the video is recommended for the user only by taking the behavior of paying attention to the video after the user browses the video as a parameter, but after ignoring the author of the video, the user may not pay attention to the page to browse the video. For example: after the user pays attention to the author of the favorite video, the user finds other videos that do not like the author. Thus, in this case, the user may not be interested in the page to browse the author's video and operate on the author's video. Therefore, the existing recommendation scheme does not pay attention to the behavior of the user after paying attention to the behavior, so that the effective conversion from the video browsing of the recommended page to the video browsing of the concerned page by the user is low, and the accuracy of recommending the multimedia resources is reduced.
In view of the above, the present disclosure provides a multimedia resource recommendation method, which, when recommending a video to a user, not only needs to consider the interactive operation (such as clicking, praise, forwarding, and focusing) on the video after the user browses the video, but also needs to consider the interactive operation (such as clicking, praise, forwarding, and other behaviors) on other videos of an author after the author of the video is focused on by the user. Therefore, the accuracy of recommending the multimedia resources to the target account is further improved by predicting the first parameter set of the target account for the video and the second parameter set of the target account for other videos of the video author, and obtaining the screening parameters according to the first parameter set and the second parameter set to generate a more accurate recommendation result.
For the convenience of understanding, the technical solutions provided by the present disclosure are further described below with reference to the accompanying drawings.
Fig. 2 is a flowchart illustrating a multimedia resource recommendation method according to an exemplary embodiment, as shown in fig. 2, including the following steps.
In step S21, acquiring a candidate multimedia resource set of the target account; and the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account.
The multimedia resources comprise multimedia information such as music, video, news and the like.
The candidate multimedia resource set is obtained by roughly ordering all videos through big data. For example: and screening out associated accounts similar to the account information of the target account from the resource library according to the account information of the target account (such as information of gender, age, region, browsed multimedia resources and the like of the user), and selecting a candidate multimedia resource set from the multimedia resources browsed by the associated accounts.
In step S22, generating a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, where the first prediction parameter set is used to record each prediction parameter for the target account to perform an interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account.
In the embodiment of the present disclosure, after the candidate multimedia resource set is obtained, the screening parameter of each multimedia resource is determined for each multimedia resource in the set. The screening parameters are composed of two parts, namely a first prediction parameter set and a second prediction parameter set.
Wherein the first set of prediction parameters includes prediction parameters of the target account's interoperation with the multimedia resource, such as: click rate, attention rate, approval rate, forwarding rate and the like. The second prediction parameter set includes prediction parameters of the target account for the interactive operation of other multimedia resources published by the publishing account after paying attention to the publishing account of the multimedia resource, for example: click rate, like rate and forward rate.
In the embodiment of the present disclosure, two prediction models may be established to predict the first prediction parameter set and the second prediction parameter set respectively. The first prediction model for predicting the first prediction parameter set takes multimedia resources as input objects and outputs and obtains all prediction parameters of the multimedia resources relative to a target account; and the second prediction model for predicting the second prediction parameter set takes the release account as an input object and outputs and obtains each prediction parameter of other multimedia resources of the target account to the release account. How to train on these two models is further explained below:
and aiming at the first pre-estimation model, acquiring a multimedia resource sample during training, wherein the multimedia resource sample comprises a user side characteristic and a content side characteristic. The user-side characteristics comprise a click list (the click list contains account information for clicking the multimedia resource, such as account identification, user gender, age and the like), a praise list (the praise list contains account information praised to the multimedia resource), an attention list (the attention list contains account information of a release account concerned with the multimedia resource) and a forwarding list (the forwarding list contains account information for forwarding the multimedia resource). The content-side features include: and issuing characteristics of account information, click rate, approval rate, attention rate, forwarding rate and the like of the account. And the multimedia resource sample has a label, and the label is the information of whether the multimedia resource sample is clicked, approved, concerned or not, forwarded or not and the like. For example: if the label of the multimedia resource sample is click, no praise, no concern and no forward, the label is a positive sample of click, a negative sample of praise, a negative sample of concern and a negative sample of forward.
And training the first estimation model according to the multimedia resource sample and the label of the multimedia resource sample, and adjusting parameters in the estimation model to enable the error between the output result and the label to be within a preset range.
And aiming at the second pre-estimation model, acquiring multimedia resource samples on the concerned page during training. Wherein the multimedia sample comprises user-side features and author-side features. The user side characteristics comprise information such as a click list, a praise list, a forwarding list and the like of the multimedia resources issued by a click author; the characteristics on the author side comprise information of author fan number, author activeness, author sex, age and the like.
In the embodiment of the disclosure, the operations performed by the user on the recommendation page and the attention page are recorded separately. For example: if a user performs operations such as approval and forwarding on one video, approves the video on a recommended page and forwards the video on a concerned page, the user information of the user can only be found in an approve list in the recommended page, and the user cannot be found in a forwarding list.
And the multimedia resource sample has a label, and the label is the information of whether the multimedia resource sample is clicked, approved or not, forwarded or not and the like. For example: if the label of the multimedia resource sample is click, no praise, no attention and no forwarding, the label is a positive sample of click, a negative sample of praise and a negative sample of forwarding.
And training the second estimation model according to the multimedia resource sample and the label of the multimedia resource sample, and adjusting parameters in the estimation model to enable the error between the output result and the label to be within a preset range.
It should be noted that the difference between the first prediction model and the second prediction model is the selection of training samples and the selection of features. The first pre-estimation model is a sample selected on a recommendation page, and the selected characteristics are user side characteristics and content side characteristics; the second pre-estimation model is a sample selected on the concerned page, and the selected characteristics are user side characteristics and author side characteristics.
After the trained second estimation model is obtained, extracting the characteristics of the account information of the release account to obtain the characteristics of the release account; and extracting the characteristics of the account information of the associated account set having the association relation with the release account to obtain the associated account characteristics.
The associated account set comprises at least two associated accounts, the account information of the associated account set comprises account information of each associated account, and interactive operation executed by each associated account on the multimedia resource of the publishing account.
Determining the second prediction parameter set of the interactive operation performed by the target account on the issuing account according to the issuing account characteristic and the associated account characteristic.
In the embodiment of the disclosure, the user side characteristic and the author side characteristic are input into the second pre-estimation model, and a second prediction parameter set is obtained by matrix calculation and the calculation result through an activation function.
In step S23, a screening parameter of the candidate multimedia resource is obtained based on the first prediction parameter set and the second prediction parameter set.
And after a first prediction parameter set and a second prediction parameter set of each multimedia resource are obtained through the two prediction models, screening parameters of each multimedia resource are obtained through processing. The method can be specifically implemented as follows:
and performing fusion processing on each prediction parameter in the second prediction parameter set to obtain a correction parameter.
In the embodiment of the present disclosure, the correction parameter is obtained by performing weighted summation on the numerical value of each prediction parameter by using a preset weighting coefficient.
And correcting the prediction parameters in the first prediction parameter set according to the correction parameters to obtain the screening parameters of the candidate multimedia resources.
In the embodiment of the present disclosure, the correction parameters may be used to correct the prediction parameters in the first prediction parameter set by the following three methods.
The first method comprises the following steps:
multiplying the prediction parameters associated with the correction parameters in the first prediction parameter set by the correction parameters to obtain corrected prediction parameters; the prediction parameter associated with the correction parameter is a prediction parameter for the target account to perform a specific operation on the publishing account through the candidate multimedia resource, wherein the specific operation is one of the interactive operations;
and carrying out weighted summation calculation on the corrected prediction parameters and other prediction parameters in the first prediction parameter set according to preset coefficients to obtain screening parameters of the candidate multimedia resources.
In the embodiment of the present disclosure, since the modification parameter is a parameter for performing an interactive operation on other multimedia resources of the publishing account after the target account pays attention to the publishing account, when performing the modification, the modification parameter may be used to modify the attention parameter in the first prediction parameter set. And multiplying the correction parameters by the attention parameters to obtain attention correction parameters, and then carrying out weighted summation on the attention correction parameters, the click parameters, the like parameters and the forwarding parameters in the first prediction parameter set by using preset weighting coefficients to obtain screening parameters. Such as: the screening parameter is a click parameter + b positive parameter + c positive parameter + d positive parameter. Wherein a, b, c and d are weighting coefficients.
And the second method comprises the following steps:
and performing weighted summation calculation on each prediction parameter in the first prediction parameter set and the correction parameter according to a preset coefficient to obtain the screening parameters of the candidate multimedia resources.
In the embodiment of the disclosure, the correction parameter can be directly used as a prediction parameter alone to perform weighted summation calculation with other prediction parameters, so as to obtain the screening parameter. Such as: the screening parameter is a click parameter + b positive parameter + c positive parameter + d positive parameter + e positive parameter. Wherein a, b, c, d and e are all weighting coefficients.
And the third is that:
carrying out feature extraction on the correction parameters to obtain correction features; performing feature extraction on each prediction parameter in the first prediction parameter set to obtain the feature of each prediction parameter;
and performing linear calculation on the characteristics of each prediction parameter and the correction characteristics to obtain the screening parameters of the candidate multimedia resources.
In the embodiment of the application, after the correction parameters and the first prediction parameter set are obtained, the prediction parameters are sequenced through a reordering model to obtain the screening parameters.
In this way, the first prediction parameter set is corrected through the correction parameters obtained by fusing the second prediction parameter set, so that the final screening score can also pay attention to other multimedia resources of the release account after the target account pays attention to the release account, and the obtained screening score is more comprehensive, and the quality of the multimedia resources recommended according to the screening score is higher.
In step S24, according to the filtering parameters, the multimedia resources are filtered from the candidate multimedia resource set and recommended to the target account.
In the embodiment of the disclosure, after the screening parameters of each multimedia resource are obtained, the multimedia resources with the preset number of the screening parameters are recommended.
In one embodiment, a predetermined number of multimedia assets may be selected from a high to a low selection parameter. For example: the candidate multimedia resource library has 10 multimedia resources in total, 5 multimedia resources are selected for recommendation, and the first 5 multimedia resources with the highest screening parameter value can be used as the multimedia resources to be recommended according to the screening parameters of the multimedia resources.
In one embodiment, a threshold value is set for the screening parameter of the multimedia resource, and the multimedia resource with the screening parameter higher than the threshold value is taken as the multimedia resource to be recommended. For example: the candidate multimedia resource library has 10 multimedia resources in total, the set threshold is 85, and if the screening parameter is higher than 85, the 4 multimedia resources are taken as the multimedia resources to be recommended.
Therefore, the quality of recommending the multimedia resources to the target account can be improved and the user retention can be effectively improved by predicting the first parameter set of the target account to the video and the second parameter set of the target account to other videos of the video author and obtaining the screening parameters according to the first parameter set and the second parameter set.
However, when recommending multimedia resources to a target account, the user viewing experience is poor due to the consistent types of the multimedia resources to be recommended, and in order to solve this problem, further screening of the multimedia resources to be recommended is required, which may be specifically implemented as:
determining multimedia resources to be recommended from the candidate multimedia resource set according to the screening parameters;
if the number of the same type of multimedia resources in the determined multimedia resources to be recommended is greater than a preset threshold value, screening the multimedia resources in the same type of which the number is greater than the preset threshold value according to the screening parameters; the types of the multimedia resources are obtained by classifying the contents of the multimedia resources;
in one embodiment, the preset threshold is set to 4, if 10 determined multimedia resources to be recommended exist, and 5 of the determined multimedia resources are of the same type, the 5 multimedia resources are further screened to obtain 4 multimedia resources, and the 4 multimedia resources and the other 5 multimedia resources are recommended to the target account.
And recommending the screened multimedia resources and other multimedia resources to be recommended to the target account.
Therefore, the multimedia resources to be recommended are further screened according to the multimedia resource types, and the watching experience of the user can be improved.
Based on the same inventive concept, the disclosure also provides a multimedia resource recommendation device. Fig. 3 is a schematic diagram of a multimedia resource recommendation apparatus according to the present disclosure. The device includes:
an acquiring unit 301 configured to perform acquiring a candidate multimedia resource set of a target account; the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account;
a prediction parameter set determining unit 302 configured to perform generation of a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, where the first prediction parameter set is used to record each prediction parameter of the target account performing an interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account;
a screening parameter determining unit 303 configured to perform screening parameters of the candidate multimedia resources based on the first prediction parameter set and the second prediction parameter set;
and the recommending unit 304 is configured to perform screening of multimedia resources from the candidate multimedia resource set according to the screening parameters to recommend to the target account.
In one possible implementation, the determining the filtering parameters unit 303 includes:
a fusion subunit, configured to perform fusion processing on each prediction parameter in the second prediction parameter set, so as to obtain a correction parameter;
and the correcting subunit is configured to correct the prediction parameters in the first prediction parameter set according to the correction parameters, so as to obtain the screening parameters of the candidate multimedia resources.
In one possible implementation, the correction subunit includes:
a first correction subunit configured to perform multiplication of a prediction parameter associated with the correction parameter in the first prediction parameter set and the correction parameter to obtain a corrected prediction parameter; the prediction parameter associated with the correction parameter is a prediction parameter for the target account to perform a specific operation on the publishing account through the candidate multimedia resource, wherein the specific operation is one of the interactive operations;
and the screening parameter determining subunit is configured to perform weighted summation calculation on the corrected prediction parameters and other prediction parameters in the first prediction parameter set according to preset coefficients to obtain screening parameters of the candidate multimedia resources.
In one possible implementation, the correction subunit includes:
and the second correction subunit is configured to perform weighted summation calculation on each prediction parameter in the first prediction parameter set and the correction parameter according to a preset coefficient to obtain a screening parameter of the candidate multimedia resource.
In one possible implementation, the correction subunit includes:
a first feature extraction subunit configured to perform feature extraction on the correction parameter to obtain a correction feature;
a second feature extraction subunit, configured to perform feature extraction on each prediction parameter in the first prediction parameter set to obtain a feature of each prediction parameter;
and the third correction subunit is configured to perform linear calculation on the characteristics of the prediction parameters and the correction characteristics to obtain the screening parameters of the candidate multimedia resources.
In a possible implementation manner, the second prediction parameter set corresponding to the candidate multimedia resource is determined by:
the first feature extraction unit is configured to perform feature extraction on the account information of the release account to obtain release account features;
the second feature extraction unit is configured to perform feature extraction on account information of an associated account set having an association relation with the release account to obtain associated account features; the associated account set comprises at least two associated accounts, the account information of the associated account set comprises account information of each associated account, and interactive operation executed by each associated account on the multimedia resource of the release account;
a determine second prediction parameter set unit configured to perform the second prediction parameter set that determines the interaction performed by the target account with the posting account according to the posting account characteristic and the associated account characteristic.
In one possible implementation manner, the recommending unit 303 includes:
a multimedia resource determining subunit configured to perform determining, according to the filtering parameter, a multimedia resource to be recommended from the candidate multimedia resource set;
the screening subunit is configured to perform screening on the multimedia resources in the same type of which the number is greater than a preset threshold value according to the screening parameters if the number of the multimedia resources in the same type in the determined multimedia resources to be recommended is greater than the preset threshold value; the types of the multimedia resources are obtained by classifying the contents of the multimedia resources;
and the recommending subunit is configured to recommend the filtered multimedia resources and other multimedia resources to be recommended to the target account.
As shown in fig. 4, based on the same technical concept, the embodiment of the present disclosure also provides an electronic device 40, which may include a memory 401 and a processor 402.
The memory 401 is used for storing computer programs executed by the processor 402. The memory 401 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to use of the task management device, and the like. The processor 402 may be a Central Processing Unit (CPU), a digital processing unit, or the like. The specific connection medium between the memory 401 and the processor 402 is not limited in the embodiments of the present disclosure. In fig. 4, the memory 401 and the processor 402 are connected by a bus 403, the bus 403 is represented by a thick line in fig. 4, and the connection manner between other components is merely illustrative and not limited. The bus 403 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. 4, but this does not indicate only one bus or one type of bus.
The memory 401 may be a volatile memory (volatile memory), such as a random-access memory (RAM); the memory 401 may also be a non-volatile memory (non-volatile memory) such as, but not limited to, a read-only memory (rom), a flash memory (flash memory), a Hard Disk Drive (HDD) or a solid-state drive (SSD), or the memory 401 may be any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. The memory 401 may be a combination of the above memories.
A processor 402 for executing the method performed by the device in the embodiment shown in fig. 1 when invoking the computer program stored in said memory 401.
In some possible embodiments, various aspects of the methods provided by the present disclosure may also be implemented in the form of a program product including program code for causing a computer device to perform the steps of the methods according to various exemplary embodiments of the present disclosure described above in this specification when the program product is run on the computer device, for example, the computer device may perform the methods performed by the devices in the embodiments shown in fig. 1-2.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
While preferred embodiments of the present disclosure have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims (10)
1. A method for recommending multimedia resources, the method comprising:
acquiring a candidate multimedia resource set of a target account; the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account;
generating a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, wherein the first prediction parameter set is used for recording each prediction parameter of the target account for performing interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account;
obtaining screening parameters of the candidate multimedia resources based on the first prediction parameter set and the second prediction parameter set;
and screening multimedia resources from the candidate multimedia resource set according to the screening parameters and recommending the multimedia resources to the target account.
2. The method of claim 1, wherein the obtaining the screening parameters of the candidate multimedia resources based on the first prediction parameter set and the second prediction parameter set comprises:
fusing all the prediction parameters in the second prediction parameter set to obtain correction parameters;
and correcting the prediction parameters in the first prediction parameter set according to the correction parameters to obtain the screening parameters of the candidate multimedia resources.
3. The method of claim 2, wherein the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources comprises:
multiplying the prediction parameters associated with the correction parameters in the first prediction parameter set by the correction parameters to obtain corrected prediction parameters; the prediction parameter associated with the correction parameter is a prediction parameter for the target account to perform a specific operation on the publishing account through the candidate multimedia resource, wherein the specific operation is one of the interactive operations;
and carrying out weighted summation calculation on the corrected prediction parameters and other prediction parameters in the first prediction parameter set according to preset coefficients to obtain screening parameters of the candidate multimedia resources.
4. The method of claim 2, wherein the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources comprises:
and performing weighted summation calculation on each prediction parameter in the first prediction parameter set and the correction parameter according to a preset coefficient to obtain the screening parameters of the candidate multimedia resources.
5. The method of claim 2, wherein the modifying the prediction parameters in the first prediction parameter set according to the modification parameters to obtain the screening parameters of the candidate multimedia resources comprises:
carrying out feature extraction on the correction parameters to obtain correction features; and the number of the first and second groups,
performing feature extraction on each prediction parameter in the first prediction parameter set to obtain the feature of each prediction parameter;
and performing linear calculation on the characteristics of each prediction parameter and the correction characteristics to obtain the screening parameters of the candidate multimedia resources.
6. The method of claim 1, wherein the second prediction parameter set corresponding to the candidate multimedia resource is determined by:
performing feature extraction on the account information of the release account to obtain release account features; and the number of the first and second groups,
carrying out feature extraction on account information of an associated account set having an association relation with the release account to obtain associated account features; the associated account set comprises at least two associated accounts, the account information of the associated account set comprises account information of each associated account, and interactive operation executed by each associated account on the multimedia resource of the release account;
determining the second prediction parameter set of the interactive operation performed by the target account on the issuing account according to the issuing account characteristic and the associated account characteristic.
7. The method for recommending multimedia resources according to claim 1, wherein said filtering multimedia resources from the candidate multimedia resource set to recommend to the target account according to the filtering parameters comprises:
determining multimedia resources to be recommended from the candidate multimedia resource set according to the screening parameters;
if the number of the same type of multimedia resources in the determined multimedia resources to be recommended is greater than a preset threshold value, screening the multimedia resources in the same type of which the number is greater than the preset threshold value according to the screening parameters; the types of the multimedia resources are obtained by classifying the contents of the multimedia resources;
and recommending the screened multimedia resources and other multimedia resources to be recommended to the target account.
8. A multimedia resource recommendation apparatus, comprising:
an acquisition unit configured to perform acquisition of a set of candidate multimedia resources of a target account; the candidate multimedia resource set comprises at least two candidate multimedia resources for recommending to the target account;
a prediction parameter set determining unit configured to perform generation of a first prediction parameter set and a second prediction parameter set corresponding to the candidate multimedia resource, where the first prediction parameter set is used to record each prediction parameter of the target account performing an interactive operation on the candidate multimedia resource; the second prediction parameter set is used for recording each prediction parameter for the target account to perform interactive operation on other multimedia resources published by the publishing account after the target account establishes a social relationship with the publishing account of the candidate multimedia resource; wherein no social relationship is established between the target account and the posting account;
a screening parameter determining unit configured to perform screening parameter obtaining on the candidate multimedia resource based on the first prediction parameter set and the second prediction parameter set;
and the recommending unit is configured to perform screening of multimedia resources from the candidate multimedia resource set according to the screening parameters and recommend the multimedia resources to the target account.
9. An electronic device, comprising:
a processor;
a memory for storing the processor-executable instructions;
wherein the processor is configured to execute the instructions to implement the multimedia resource recommendation method of any of claims 1-7.
10. A storage medium, wherein instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the multimedia asset recommendation method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010378667.0A CN113626679B (en) | 2020-05-07 | 2020-05-07 | Multimedia resource recommendation method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010378667.0A CN113626679B (en) | 2020-05-07 | 2020-05-07 | Multimedia resource recommendation method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113626679A true CN113626679A (en) | 2021-11-09 |
| CN113626679B CN113626679B (en) | 2023-10-17 |
Family
ID=78376940
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010378667.0A Active CN113626679B (en) | 2020-05-07 | 2020-05-07 | Multimedia resource recommendation method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113626679B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114547417A (en) * | 2022-02-25 | 2022-05-27 | 北京百度网讯科技有限公司 | Media resource ordering method and electronic equipment |
| CN114819000A (en) * | 2022-06-29 | 2022-07-29 | 北京达佳互联信息技术有限公司 | Feedback information estimation model training method and device and electronic equipment |
| CN115065677A (en) * | 2022-07-22 | 2022-09-16 | 北京达佳互联信息技术有限公司 | Media resource acquisition method, device, electronic equipment, storage medium and product |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106464682A (en) * | 2014-05-22 | 2017-02-22 | 谷歌公司 | Using status of sign-on to online services for content item recommendations |
| US20170171334A1 (en) * | 2015-12-14 | 2017-06-15 | Le Holdings (Beijing) Co., Ltd. | Single-account multiple-preference recommendation method for video website and electronic device |
| CN107851277A (en) * | 2015-08-10 | 2018-03-27 | 谷歌有限责任公司 | The consistent and personalized social media content of privacy is shared to recommend |
| WO2018059176A1 (en) * | 2016-09-27 | 2018-04-05 | 腾讯科技（深圳）有限公司 | Method and apparatus for generating targeted label and storage medium |
| WO2018090793A1 (en) * | 2016-11-18 | 2018-05-24 | 腾讯科技（深圳）有限公司 | Multimedia recommendation method and device |
-
2020
- 2020-05-07 CN CN202010378667.0A patent/CN113626679B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106464682A (en) * | 2014-05-22 | 2017-02-22 | 谷歌公司 | Using status of sign-on to online services for content item recommendations |
| CN107851277A (en) * | 2015-08-10 | 2018-03-27 | 谷歌有限责任公司 | The consistent and personalized social media content of privacy is shared to recommend |
| US20170171334A1 (en) * | 2015-12-14 | 2017-06-15 | Le Holdings (Beijing) Co., Ltd. | Single-account multiple-preference recommendation method for video website and electronic device |
| WO2018059176A1 (en) * | 2016-09-27 | 2018-04-05 | 腾讯科技（深圳）有限公司 | Method and apparatus for generating targeted label and storage medium |
| WO2018090793A1 (en) * | 2016-11-18 | 2018-05-24 | 腾讯科技（深圳）有限公司 | Multimedia recommendation method and device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114547417A (en) * | 2022-02-25 | 2022-05-27 | 北京百度网讯科技有限公司 | Media resource ordering method and electronic equipment |
| CN114819000A (en) * | 2022-06-29 | 2022-07-29 | 北京达佳互联信息技术有限公司 | Feedback information estimation model training method and device and electronic equipment |
| CN114819000B (en) * | 2022-06-29 | 2022-10-21 | 北京达佳互联信息技术有限公司 | Feedback information estimation model training method and device and electronic equipment |
| CN115065677A (en) * | 2022-07-22 | 2022-09-16 | 北京达佳互联信息技术有限公司 | Media resource acquisition method, device, electronic equipment, storage medium and product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113626679B (en) | 2023-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106649316B (en) | Video pushing method and device | |
| US9684656B2 (en) | Creating personalized and continuous playlists for a content sharing platform based on user history | |
| CN110149540B (en) | Recommendation processing method and device for multimedia resources, terminal and readable medium | |
| CN113626679B (en) | Multimedia resource recommendation method, device and storage medium | |
| CN107885796B (en) | Information recommendation method, device and equipment | |
| CN110674408B (en) | Service platform, and real-time generation method and device of training sample | |
| CN102262647B (en) | Signal conditioning package, information processing method and program | |
| US11748408B2 (en) | Analyzing user searches of verbal media content | |
| CN108509417B (en) | Title generation method and device, storage medium and server | |
| US10796339B2 (en) | Detecting expired content within slots in a user interface | |
| CN109255037B (en) | Method and apparatus for outputting information | |
| CN113688313A (en) | Training method of prediction model, information pushing method and device | |
| CN112987996B (en) | Information display method, information display device, electronic equipment and computer readable storage medium | |
| CN112883257B (en) | Behavior sequence data processing method and device, electronic equipment and storage medium | |
| CN113656681B (en) | Object evaluation method, device, equipment and storage medium | |
| US9424357B1 (en) | Predictive page loading based on text entry and search term suggestions | |
| CN112040339A (en) | Method and device for making video data, computer equipment and storage medium | |
| CN110717788A (en) | Target user screening method and device | |
| US8903817B1 (en) | Determining search relevance from user feedback | |
| CN116738060A (en) | Content generation method and device and electronic equipment | |
| CN115129922A (en) | Search term generation method, model training method, medium, device and equipment | |
| CN114331602A (en) | Model training method based on transfer learning, information recommendation method and device | |
| CN113343085A (en) | Information recommendation method and device, storage medium and electronic equipment | |
| CN111400512A (en) | Method and device for screening multimedia resources | |
| CN114491264A (en) | Recommendation model training method, medium, device and computing equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |