WO2019024670A1 - 多媒体资源的推荐方法及装置 - Google Patents

多媒体资源的推荐方法及装置 Download PDF

Info

Publication number
WO2019024670A1
WO2019024670A1 PCT/CN2018/095586 CN2018095586W WO2019024670A1 WO 2019024670 A1 WO2019024670 A1 WO 2019024670A1 CN 2018095586 W CN2018095586 W CN 2018095586W WO 2019024670 A1 WO2019024670 A1 WO 2019024670A1
Authority
WO
WIPO (PCT)
Prior art keywords
multimedia resource
multimedia
candidate
resources
recommended
Prior art date
Application number
PCT/CN2018/095586
Other languages
English (en)
French (fr)
Inventor
滕飞
Original Assignee
优酷信息技术（北京）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 优酷信息技术（北京）有限公司 filed Critical 优酷信息技术（北京）有限公司
Publication of WO2019024670A1 publication Critical patent/WO2019024670A1/zh

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
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/48—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • 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

Definitions

  • the present disclosure relates to the field of multimedia technologies, and in particular, to a method and an apparatus for recommending multimedia resources.
  • the present disclosure proposes a method and apparatus for recommending multimedia resources to improve the recommendation effect of multimedia resources.
  • a method for recommending a multimedia resource including:
  • determining the multimedia resource to be selected includes:
  • determining the multimedia resource to be selected includes:
  • the multimedia resource that is not viewed by the target user in the selectable multimedia resource is determined as a multimedia resource to be selected.
  • determining the multimedia resource to be selected includes:
  • the multimedia resource to be selected is determined from each multimedia resource of the multimedia resource library according to one or more of the search quantity, the click volume and the click rate of each multimedia resource in the multimedia resource library.
  • determining an estimated amount of clicks of each of the candidate multimedia resources including:
  • For each of the second types of candidate multimedia resources searching for the first type of candidate multimedia resources that match the second type of candidate multimedia resources, and calculating the second type of candidate multimedia resources and the matched first
  • the matching degree of the candidate multimedia resource is calculated, and the estimated amount of the click amount of the second type of candidate multimedia resource is calculated according to the matching degree and the estimated click value of the matched first-class multimedia resource to be selected.
  • the multimedia resources to be recommended are determined from each of the to-be-selected multimedia resources according to the recommended scenario and the estimated amount of clicks of the selected multimedia resources, including:
  • the multimedia resource recommendation is performed to the target user according to the multimedia resource to be recommended, including:
  • the recommendation result is recommended to the target user.
  • a device for recommending a multimedia resource including:
  • a first determining module configured to determine a plurality of candidate multimedia resources
  • a second determining module configured to determine an estimated amount of clicks of each of the to-be-selected multimedia resources
  • the third determining module determines, according to the recommended scenario and the estimated amount of clicks of the selected multimedia resources, the multimedia resources to be recommended from each of the to-be-selected multimedia resources;
  • the recommendation module is configured to perform multimedia resource recommendation to the target user according to the multimedia resource to be recommended.
  • the first determining module includes:
  • the first determining submodule is configured to determine, according to the label of the target user, and the label of each multimedia resource in the multimedia resource library, the multimedia resource to be selected from the multimedia resources of the multimedia resource library.
  • the first determining module includes:
  • a second determining submodule configured to determine, according to user behavior data of each user in the user cluster to which the target user belongs, an optional multimedia resource corresponding to the user cluster;
  • a third determining submodule configured to determine, as the candidate multimedia resource, the multimedia resource that is not viewed by the target user in the optional multimedia resource.
  • the first determining module includes:
  • a fourth determining submodule configured to determine, according to one or more of a search quantity, a click quantity, and a click rate of each multimedia resource in the multimedia resource library, the multimedia resource to be selected from the multimedia resources of the multimedia resource library.
  • the second determining module includes:
  • a sub-module configured to divide the candidate multimedia resource into a first-class candidate multimedia resource and a second-class candidate multimedia resource
  • a fifth determining submodule configured to determine, according to historical click data of the first type of candidate multimedia resources, a click amount estimation of the first type of multimedia resource to be selected for each of the first type of candidate multimedia resources value;
  • a sixth determining submodule configured to search for a first type of candidate multimedia resource that matches the second type of candidate multimedia resource for each of the second type of candidate multimedia resources, and calculate the second type of candidate to be selected The matching degree between the multimedia resource and the matched first-class candidate multimedia resource, and calculating the second-class candidate multimedia resource according to the matching degree and the matched predicted value of the first-class candidate multimedia resource Estimated traffic.
  • the third determining module includes:
  • Obtaining a sub-module configured to obtain a recommended scenario corresponding to each recommendation window in the recommendation page of the multimedia resource
  • a seventh determining sub-module configured to determine, from each of the candidate multimedia resources, a to-be-recommended for each recommended window according to a recommended scenario corresponding to each recommended window, and a predicted amount of clicks of each of the to-be-selected multimedia resources Multimedia resources.
  • the recommendation module includes:
  • An eighth determining submodule configured to determine a recommendation result from the multimedia resources to be recommended according to at least one of an uploader of the multimedia resource to be recommended, an associated channel, and a label;
  • a recommendation submodule is used to recommend the recommendation result to the target user.
  • a recommendation device for a multimedia resource comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to perform the above method.
  • a non-transitory computer readable storage medium having stored thereon computer program instructions, wherein the computer program instructions are implemented by a processor to implement the above method.
  • the method and device for recommending multimedia resources of various aspects of the present disclosure determine a plurality of to-be-selected multimedia resources, determine an estimated amount of clicks of each candidate multimedia resource, and estimate the click amount according to the recommended scenario and each candidate multimedia resource. Determining the multimedia resource to be recommended from each of the to-be-selected multimedia resources, and performing multimedia resource recommendation to the target user according to the multimedia resource to be recommended, thereby improving the recommendation effect of the multimedia resource.
  • FIG. 1 illustrates a flow chart of a method of recommending a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 2 illustrates an exemplary flowchart of a method S12 of recommending a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 3 illustrates an exemplary flowchart of a method S127 of recommending a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 4 illustrates another exemplary flowchart of a recommendation method step S121 of a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 5 illustrates an exemplary flowchart of step S122 of a method for recommending multimedia resources according to an embodiment of the present disclosure.
  • FIG. 6 shows an exemplary flowchart of step S123 of the recommendation method of the multimedia resource according to an embodiment of the present disclosure.
  • FIG. 7 illustrates a block diagram of a recommendation device for a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 8 illustrates an exemplary block diagram of a recommendation device for a multimedia resource according to an embodiment of the present disclosure.
  • FIG. 9 is a block diagram of an apparatus 1900 for recommending multimedia resources, according to an exemplary embodiment.
  • FIG. 1 illustrates a flow chart of a method of recommending a multimedia resource according to an embodiment of the present disclosure.
  • the method can be applied to a server, which is not limited herein. As shown in FIG. 1, the method includes steps S11 to S14.
  • step S11 a plurality of candidate multimedia resources are determined.
  • the multimedia may be a combination of a plurality of media, including, for example, text, sound, and images.
  • the multimedia resource may be one or more of video and audio, and is not limited herein.
  • determining the to-be-selected multimedia resource may include: determining, according to the label of the target user, and the label of each multimedia resource in the multimedia resource library, the multimedia resource to be selected from the multimedia resources of the multimedia resource library.
  • the tag of the target user may be determined based on user behavior data of the target user.
  • the user behavior data of the target user may include, but is not limited to, at least one of the following: a multimedia resource viewed by the target user, a multimedia resource of the target user comment, a multimedia resource of the target user publishing the barrage, a multimedia resource purchased by the target user, and a subscription of the target user.
  • the label of the target user may be determined according to one or both of the title and the profile of each multimedia resource in the user behavior data of the target user.
  • a multimedia resource library may refer to a database that stores all or a portion of multimedia resources in a video website.
  • the tag of the multimedia resource may be determined according to one or both of the title and profile of the multimedia resource.
  • the title and the introduction of the multimedia resource may be segmented, and the label of the multimedia resource is determined according to the result of the word segmentation.
  • the similarity between the label of the target user and the label of each multimedia resource may be calculated, and the candidate multimedia resource may be determined according to the similarity between the label of the target user and the label of each multimedia resource.
  • the multimedia resource with the similarity of the label of the target user that is greater than the first threshold may be determined as the multimedia resource to be selected.
  • the N multimedia resources with the highest degree of similarity with the tag of the target user may be determined as the candidate multimedia resource, where N is a positive integer.
  • this implementation can be implemented based on a content recommendation model.
  • the implementation determines the candidate multimedia resource from each multimedia resource of the multimedia resource library according to the label of the target user and the label of each multimedia resource in the multimedia resource library, and the determined candidate multimedia resource can reflect the real interest of the target user. , which helps to improve the accuracy and effectiveness of multimedia resource recommendations.
  • determining the to-be-selected multimedia resource may include: determining, according to user behavior data of each user in the user cluster to which the target user belongs, an optional multimedia resource corresponding to the user cluster; The multimedia resource that is not viewed by the target user is determined to be a multimedia resource to be selected.
  • the optional multimedia resources corresponding to the user cluster are C1, C2, C3, C4, C5, and C6. If the multimedia resource that the target user has viewed is C3, C1, C2, C4, C5, and C6 may be determined as Selected multimedia resources.
  • multiple users with similar user behavior data can be attributed to the same user cluster.
  • the user behavior data of the multiple users are similar, it may indicate that the preferences of the multiple users viewing the multimedia resources are similar.
  • multiple users with the same or similar tags can be grouped into the same user cluster.
  • the target user may be added to the user cluster of interest to the target user based on the request of the target user.
  • determining the optional multimedia resource corresponding to the user cluster according to the user behavior data of each user in the user cluster to which the target user belongs may include: multimedia in the user behavior data of each user in the user cluster.
  • the resource is determined to be an optional multimedia resource corresponding to the user cluster.
  • this implementation may be implemented based on a collaborative filtering model of multimedia resources.
  • the implementation determines the optional multimedia resource corresponding to the user cluster according to the user behavior data of each user in the user cluster to which the target user belongs, and determines the multimedia resource that is not viewed by the target user in the optional multimedia resource as the candidate multimedia resource, This can mine the multimedia resources that the target user may be interested in through the user behavior data of the user cluster, thereby helping to improve the accuracy and effect of the multimedia resource recommendation.
  • determining the to-be-selected multimedia resource may include: searching for each of the multimedia resource pools according to one or more of a search quantity, a click volume, and a click rate of each multimedia resource in the multimedia resource library.
  • the multimedia resource to be selected is determined in the multimedia resource.
  • the hotspot score of each multimedia resource may be determined according to one or more of the search quantity, the click volume, and the click rate of each multimedia resource in the multimedia resource library in a specified time period, and The M multimedia resources with the highest hotspot score are determined as the candidate multimedia resources, where M is a positive integer.
  • the specified time period can be 3 days.
  • the hotspot score S r of a certain multimedia resource in the multimedia resource library in a specified time period may be calculated by using Equation 3:
  • R 1 represents a search amount of the multimedia resource in a specified time period
  • ⁇ 1 represents a weight corresponding to R 1
  • R 2 represents a click amount of the multimedia resource in a specified time period
  • ⁇ 2 represents a weight corresponding to R 2
  • R 3 represents the click rate of the multimedia resource within a specified time period
  • ⁇ 3 represents the weight corresponding to R 3 .
  • the P multimedia resources with the highest hotspot score among the multiple multimedia resource categories may be determined as the candidate multimedia resources, where P is a positive integer.
  • P is a positive integer.
  • P multimedia resources with the highest hotspot score in the movie classification P multimedia resources with the highest hotspot score in the TV drama category, P multimedia resources with the highest hotspot score in the variety classification, and P multimedia with the highest hotspot score in the sports classification.
  • the resources are respectively determined as the multimedia resources to be selected.
  • this implementation can be implemented based on a hotspot model.
  • the implementation method determines the multimedia resource to be selected from each multimedia resource of the multimedia resource library according to one or more of the search quantity, the click volume and the click rate of each multimedia resource in the multimedia resource library, thereby enabling the multimedia of the hot spot
  • the resource is determined to be a multimedia resource to be selected, thereby contributing to the improvement of the recommendation of the multimedia resource.
  • step S12 an estimated amount of clicks of each candidate multimedia resource is determined.
  • step S13 the multimedia resource to be recommended is determined from each candidate multimedia resource according to the recommended scenario and the estimated amount of clicks of the multimedia resources to be selected.
  • the scores of the candidate multimedia resources can be calculated by combining the estimated amount of the clicks of the multimedia resources to be selected, and the recommended multimedia resources can be determined as the to-be-recommended.
  • Multimedia resources The recommendation scene may include one or more of a popular recommendation, a personalized recommendation, a channel recommendation, and a festival-related recommendation.
  • the score of each candidate multimedia resource may be calculated according to the real-time search volume of the multimedia resource to be selected and the estimated amount of the clicked multimedia resource.
  • the score of each candidate multimedia resource may be calculated according to the matching degree between the tag of the multimedia resource to be recommended and the tag of the target user, and the estimated value of the click amount of the multimedia resource to be selected.
  • the score of each candidate multimedia resource may be calculated according to the matching degree of the multimedia resource to be selected and the predicted amount of the clicked multimedia resource.
  • the channel may include a TV drama channel, a news channel, a variety channel, and a sports channel.
  • the same candidate multimedia resource may have different scores on different channels.
  • the score of each candidate multimedia resource may be calculated according to the matching degree between the multimedia resource to be recommended and the current holiday, and the estimated amount of the clicked multimedia resource to be selected. For example, during the Spring Festival, multimedia resources related to the Spring Festival can be recommended to target users.
  • the number of recommended multimedia resources to be selected may also be adjusted according to an application platform.
  • the application platform may include a webpage end and an App (Application) end.
  • the multimedia resources to be recommended are determined from the respective candidate multimedia resources according to the recommended scenario and the estimated amount of the clicks of the multimedia resources to be selected, including: obtaining the recommendation in the recommendation page of the multimedia resource.
  • the recommended scenario corresponding to the window; the recommended multimedia resource for each recommendation window is determined from each candidate multimedia resource according to the recommended scenario corresponding to each recommendation window and the estimated traffic value of each candidate multimedia resource.
  • the recommendation window includes a first recommendation window, a second recommendation window, a third recommendation window, and a fourth recommendation window.
  • the recommended scene corresponding to the first recommendation window is a hot recommendation
  • the recommended scene corresponding to the second recommendation window is a personalized recommendation
  • the recommended scene corresponding to the third recommendation window is a festival-related recommendation
  • the recommended scene corresponding to the fourth recommendation window is a TV drama. Channel recommendation.
  • step S14 the multimedia resource recommendation is performed to the target user according to the multimedia resource to be recommended.
  • the multimedia resource recommendation may be performed to the target user according to the multimedia resource to be recommended, when receiving the recommendation request from the target user.
  • one of the highest-recommended multimedia resources to be recommended may be recommended to the target user.
  • a plurality of multimedia resources to be recommended with the highest score may be recommended to the target user.
  • all multimedia resources to be recommended may be recommended to the target user.
  • the performing the multimedia resource recommendation to the target user according to the multimedia resource to be recommended may include: recommending, according to at least one of an uploader, a channel, and a label of the multimedia resource to be recommended. Determining a recommendation result in the multimedia resource; recommending the recommendation result to the target user.
  • the recommendation result may include at most P multimedia resources uploaded by the same uploader to improve the diversity of the recommendation results.
  • P is a positive integer.
  • P is equal to 3.
  • the recommendation result may include at most Q multimedia resources of the same secondary channel to improve the diversity of the recommendation results.
  • Q is a positive integer.
  • Q is equal to 3.
  • the variety channel is a certain level channel
  • the Hunan variety channel is the second level channel under the level one channel.
  • the recommendation result may include at most R multimedia resources having the same label to improve the diversity of the recommendation results.
  • R is a positive integer.
  • R is equal to 3.
  • the candidate multimedia resources are determined.
  • FIG. 2 illustrates an exemplary flowchart of a method S12 of recommending a multimedia resource according to an embodiment of the present disclosure. As shown in FIG. 2, step S12 may include steps S121 to S123.
  • step S121 the to-be-selected multimedia resources are divided into a first type of candidate multimedia resources and a second type of candidate multimedia resources.
  • all the to-be-selected multimedia resources can be divided into the first type of candidate multimedia resources and the second type of candidate multimedia resources, and are respectively used for the first type of candidate multimedia resources and the second type of candidate multimedia resources. Different methods calculate their traffic estimates.
  • step S122 for each of the first type of candidate multimedia resources, the estimated amount of clicks of the first type of multimedia resources to be selected is determined according to historical click data of the first type of candidate multimedia resources.
  • step S123 for each of the second type of candidate multimedia resources, searching for the first type of candidate multimedia resources that match the second type of candidate multimedia resources, and calculating the second type of candidate multimedia resources and the matched first class.
  • the matching degree of the candidate multimedia resource is calculated, and the estimated amount of the click amount of the second type of candidate multimedia resource is calculated according to the matching degree and the estimated click value of the matched first-class multimedia resource to be selected.
  • the second type is selected to be 1 v multimedia resources, if it can find the candidate with the second type of matching multimedia resources v 1 is selected from a first class of multimedia resources to be v 2, selected from the second category to be a Multimedia Resource Calculation v The degree of matching with the first type of multimedia resource v 2 to be selected, and then calculating the estimated amount of the second type of candidate multimedia resource v 1 according to the matching degree and the estimated amount of the click of the first type of candidate multimedia resource v 2 .
  • FIG. 3 illustrates an exemplary flowchart of a method S127 of recommending a multimedia resource according to an embodiment of the present disclosure.
  • step S121 may include steps S1211 and S1212.
  • step S1211 historical click data of each candidate multimedia resource is respectively acquired.
  • step S1212 for each of the to-be-selected multimedia resources, the current click data of the multimedia resource to be selected is determined to be in a falling period according to the historical click data of the multimedia resource to be selected, and if so, the multimedia resource to be selected is determined as the first type of waiting. The multimedia resource is selected, otherwise the candidate multimedia resource is determined as the second type of candidate multimedia resource.
  • the current click volume of the multimedia resource to be selected is in a falling period, which may indicate that the amount of the clicked multimedia resource decreases with time.
  • FIG. 4 illustrates another exemplary flowchart of a recommendation method step S121 of a multimedia resource according to an embodiment of the present disclosure. As shown in FIG. 4, step S121 may include steps S1213 and S1214.
  • step S1213 the upload time of each candidate multimedia resource is respectively acquired.
  • step S1214 it is determined, for each of the to-be-selected multimedia resources, whether the time length of the uploading time of the multimedia resource to be selected is greater than a second threshold, and if so, determining the multimedia resource to be selected as the first type of candidate multimedia. Resources, otherwise the candidate multimedia resources are determined as the second type of candidate multimedia resources.
  • the second threshold may be 48 hours, which is not limited herein.
  • FIG. 5 illustrates an exemplary flowchart of step S122 of a method for recommending multimedia resources according to an embodiment of the present disclosure. As shown in FIG. 5, step S122 may include steps S1221 through S1223.
  • step S1221 historical click data of the first type of candidate multimedia resources is obtained for each of the first type of candidate multimedia resources.
  • step S1222 the time decay coefficient of the first type of candidate multimedia resources is trained according to historical click data of the first type of candidate multimedia resources.
  • step S1223 the estimated amount of clicks of the first type of candidate multimedia resources is calculated according to the historical click data and the time decay coefficient of the first type of candidate multimedia resources.
  • the training the time decay coefficient of the first type of candidate multimedia resources according to the history click data of the first type of candidate multimedia resources may include: according to the history of the first type of candidate multimedia resources i Clicking on the data, using the formula 1 to train the time decay coefficient ⁇ of the first type of candidate multimedia resource i, wherein the historical click data of the first type of candidate multimedia resource i includes the click of the first type of candidate multimedia resource i before the specified date the amount;
  • V i0 represents the click amount of the first type of candidate multimedia resource i on the specified date
  • f(0) V i0
  • Day, f(t) represents the number of hits of the first type of candidate multimedia resource i on the tth day before the specified date; when t ⁇ 0, t represents the first-t day after the specified date, and f(t) represents The estimated amount of traffic for the first-class candidate multimedia resource i on the first-t day after the specified date.
  • the historical click data of the first type of candidate multimedia resource i includes the click amount of the first type of candidate multimedia resource i on the specified date and the click amount of the first type of candidate multimedia resource i 15 days before the specified date.
  • f(1) indicates the amount of clicks of the first type of candidate multimedia resource i on the day before the specified date
  • f(2) indicates the first
  • the number of clicks of a candidate multimedia resource i on the first two days of the specified date, and so on, t 15 indicates the first 15 days of the specified date
  • f(15) indicates that the first type of candidate multimedia resource i is on the specified date. The number of clicks in the first 15 days.
  • Equation 1 When training the time decay coefficient ⁇ of the first type of candidate multimedia resource i using Equation 1, f(t), V i0 and t in Equation 1 are known for t>0, according to f(t), V I0 and t can train to obtain the time decay coefficient ⁇ of the first type of candidate multimedia resource i.
  • the time decay coefficient ⁇ can also be corrected according to the history click data of the first type of candidate multimedia resource i.
  • calculating, according to the historical click data and the time decay coefficient of the first type of candidate multimedia resources, the estimated amount of the click amount of the first type of candidate multimedia resources which may include: calculating the formula 1 after the training The estimated amount of traffic for a class of candidate multimedia resources i on the first -t day after the specified date.
  • f(-1) indicates the amount of clicks of the first type of candidate multimedia resource i on the day before the specified date.
  • the time decay coefficient ⁇ of the first type of candidate multimedia resource i for t ⁇ 0, V i0 , ⁇ and t in Equation 1 are known, so that one day after the specified date can be calculated.
  • Click volume estimate For example, using the trained formula 1 to calculate the estimated amount of the first-day candidate multimedia resource i on the first-t day after the specified date, the calculation may be: using the trained formula 1 to calculate f(-1), f(-1) is determined as the estimated amount of clicks of the first type of candidate multimedia resource i.
  • the training the time decay coefficient of the first type of candidate multimedia resources according to the history click data of the first type of candidate multimedia resources may include: according to the first type of candidate multimedia resources i
  • the historical click data is used to train the time decay coefficient ⁇ of the first type of candidate multimedia resource i using Equation 2, wherein the historical click data of the first type of candidate multimedia resource i includes the first type of candidate multimedia resource i before the specified date.
  • V i0 represents the click amount of the first type of candidate multimedia resource i on the specified date
  • g(0) lgV i0
  • g(t) represents the logarithm of the click amount of the first type of candidate multimedia resource i on the tth day before the specified date
  • t ⁇ 0 t represents the first-t day after the specified date
  • g( t) represents the logarithm of the estimated amount of clicks of the first type of candidate multimedia resource i on the first-t day after the specified date.
  • the estimated click value of the first type of candidate multimedia resources is calculated according to the historical click data and the time decay coefficient of the first type of candidate multimedia resources, including: calculating the first by using the trained formula 2 The logarithm of the estimated amount of hits for the class-t days after the specified date.
  • the logarithm of the click volume is reduced, thereby reducing the value for ease of calculation and storage.
  • the logarithm of the estimated value of the first-to-day day of the first-class candidate multimedia resource i calculated according to Equation 2 may be used as the rank of the first-class candidate multimedia resource i ( Sort) values.
  • FIG. 6 shows an exemplary flowchart of step S123 of the recommendation method of the multimedia resource according to an embodiment of the present disclosure. As shown in FIG. 6, step S123 may include steps S1231 and S1232.
  • step S1231 for each of the second type of candidate multimedia resources, the first type of candidate multimedia resources that match the second type of candidate multimedia resources are searched according to the specified information of the second type of candidate multimedia resources; the specified information includes the following: At least one item: upload time, length of time, and uploader information.
  • step S1232 the second type of candidate multimedia resources and the matched first type of candidate multimedia resources are calculated according to the specified information of the second type of candidate multimedia resources and the matching information of the matched first type of candidate multimedia resources. suitability.
  • the matching degree between the second type of candidate multimedia resources and the matched first type of candidate multimedia resources is greater than 0 and less than or equal to 1.
  • the to-be-selected multimedia resource is a video
  • the length of the candidate multimedia resource may be referred to as a video length.
  • the matching degree between the second type of candidate multimedia resources and the first type of candidate multimedia resources is compared. high.
  • the matching degree between the second type of candidate multimedia resources and the matched first type of candidate multimedia resources may be determined according to an uploading time, a length of time, and an uploader information, where the uploading time is for the matching degree.
  • the weight of the time is ⁇ 1
  • the weight of the time length is ⁇ 2 for the matching degree
  • the weight of the uploader information for the matching degree is ⁇ 3 .
  • calculating, by the matching degree, the estimated amount of the click quantity of the second type of candidate multimedia resources according to the matched estimated value of the first type of candidate multimedia resources which may include: matching The product of the degree and the estimated amount of clicks of the matched first-class candidate multimedia resources is used as the estimated amount of clicks of the second type of candidate multimedia resources.
  • FIG. 7 illustrates a block diagram of a recommendation device for a multimedia resource according to an embodiment of the present disclosure.
  • the apparatus includes: a first determining module 71, configured to determine a plurality of candidate multimedia resources; and a second determining module 72, configured to determine a click amount estimation value of each of the candidate multimedia resources;
  • the determining module 73 determines the multimedia resource to be recommended from each of the to-be-selected multimedia resources according to the recommended scenario and the estimated amount of the clicked multimedia resources, and the recommendation module 74 is configured to The multimedia resource provides multimedia resource recommendation to the target user.
  • FIG. 8 illustrates an exemplary block diagram of a recommendation device for a multimedia resource according to an embodiment of the present disclosure. As shown in Figure 8:
  • the first determining module 71 includes: a first determining submodule 711, configured to use, according to the label of the target user, and a label of each multimedia resource in the multimedia resource library, from the multimedia The multimedia resources to be selected are determined in each multimedia resource of the resource library.
  • the first determining module 71 includes: a second determining sub-module 712, configured to determine, according to user behavior data of each user in the user cluster to which the target user belongs,
  • the third determining sub-module 713 is configured to determine, as the candidate multimedia resource, the multimedia resource that is not viewed by the target user in the optional multimedia resource.
  • the first determining module 71 includes: a fourth determining submodule 714, configured to use one or more of a search volume, a click volume, and a click rate of each multimedia resource in the multimedia resource library. And determining a candidate multimedia resource from each multimedia resource of the multimedia resource library.
  • the second determining module 72 includes: a dividing sub-module 721, configured to divide the candidate multimedia resource into a first type of candidate multimedia resource and a second type of candidate multimedia resource;
  • the fifth determining sub-module 722 is configured to determine, according to historical click data of the first type of candidate multimedia resources, a click amount of the first type of multimedia resource to be selected for each of the first type of candidate multimedia resources.
  • a sixth determining sub-module 723, configured to search for a first type of candidate multimedia resource that matches the second type of candidate multimedia resource for each of the second type of candidate multimedia resources, and calculate the first The matching degree between the second type of candidate multimedia resources and the matched first type of candidate multimedia resources, and calculating the second category according to the matching degree and the matched estimated amount of the first type of candidate multimedia resources The estimated amount of traffic for the selected multimedia resource.
  • the third determining module 73 includes: an obtaining sub-module 731, configured to obtain a recommended scenario corresponding to each recommended window in the recommended page of the multimedia resource; and a seventh determining sub-module 732, configured to a recommended scenario corresponding to the window, and an estimated amount of clicks of each of the to-be-selected multimedia resources, and determining, from each of the candidate multimedia resources, a multimedia resource to be recommended for each recommended window.
  • the recommendation module 74 includes: an eighth determining sub-module 741, configured to: according to at least one of an uploader, a channel, and a label of the multimedia resource to be recommended, Determining a recommendation result in the recommended multimedia resource; a recommendation sub-module 742, configured to recommend the recommendation result to the target user.
  • the candidate multimedia resources are determined.
  • FIG. 9 is a block diagram of an apparatus 1900 for recommending multimedia resources, according to an exemplary embodiment.
  • device 1900 can be provided as a server.
  • apparatus 1900 includes a processing component 1922 that further includes one or more processors, and memory resources represented by memory 1932 for storing instructions executable by processing component 1922, such as an application.
  • An application stored in memory 1932 can include one or more modules each corresponding to a set of instructions.
  • processing component 1922 is configured to execute instructions to perform the methods described above.
  • Apparatus 1900 can also include a power supply component 1926 configured to perform power management of apparatus 1900, a wired or wireless network interface 1950 configured to connect apparatus 1900 to the network, and an input/output (I/O) interface 1958.
  • Device 1900 can operate based on an operating system stored in memory 1932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, or the like.
  • a non-transitory computer readable storage medium such as a memory 1932 comprising computer program instructions executable by processing component 1922 of apparatus 1900 to perform the above method.
  • the present disclosure can be a system, method, and/or computer program product.
  • the computer program product can comprise a computer readable storage medium having computer readable program instructions embodied thereon for causing a processor to implement various aspects of the present disclosure.
  • the computer readable storage medium can be a tangible device that can hold and store the instructions used by the instruction execution device.
  • the computer readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, for example, with instructions stored thereon A raised structure in the hole card or groove, and any suitable combination of the above.
  • a computer readable storage medium as used herein is not to be interpreted as a transient signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (eg, a light pulse through a fiber optic cable), or through a wire The electrical signal transmitted.
  • the computer readable program instructions described herein can be downloaded from a computer readable storage medium to various computing/processing devices or downloaded to an external computer or external storage device over a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and/or edge servers.
  • a network adapter card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in each computing/processing device .
  • Computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine related instructions, microcode, firmware instructions, state setting data, or in one or more programming languages.
  • Source code or object code written in any combination including object oriented programming languages such as Smalltalk, C++, etc., as well as conventional procedural programming languages such as the "C" language or similar programming languages.
  • the computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer, partly on the remote computer, or entirely on the remote computer or server. carried out.
  • the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or wide area network (WAN), or can be connected to an external computer (eg, using an Internet service provider to access the Internet) connection).
  • the customized electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), can be customized by utilizing state information of computer readable program instructions.
  • Computer readable program instructions are executed to implement various aspects of the present disclosure.
  • the computer readable program instructions can be provided to a general purpose computer, a special purpose computer, or a processor of other programmable data processing apparatus to produce a machine such that when executed by a processor of a computer or other programmable data processing apparatus Means for implementing the functions/acts specified in one or more of the blocks of the flowcharts and/or block diagrams.
  • the computer readable program instructions can also be stored in a computer readable storage medium that causes the computer, programmable data processing device, and/or other device to operate in a particular manner, such that the computer readable medium storing the instructions includes An article of manufacture that includes instructions for implementing various aspects of the functions/acts recited in one or more of the flowcharts.
  • the computer readable program instructions can also be loaded onto a computer, other programmable data processing device, or other device to perform a series of operational steps on a computer, other programmable data processing device or other device to produce a computer-implemented process.
  • instructions executed on a computer, other programmable data processing apparatus, or other device implement the functions/acts recited in one or more of the flowcharts and/or block diagrams.
  • each block in the flowchart or block diagram can represent a module, a program segment, or a portion of an instruction that includes one or more components for implementing the specified logical functions.
  • Executable instructions can also occur in a different order than those illustrated in the drawings. For example, two consecutive blocks may be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts can be implemented in a dedicated hardware-based system that performs the specified function or action. Or it can be implemented by a combination of dedicated hardware and computer instructions.

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)
  • Library & Information Science (AREA)
  • Multimedia (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本公开涉及多媒体资源的推荐方法及装置。该方法包括：确定多个待选多媒体资源；确定各个所述待选多媒体资源的点击量预估值；根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。本公开能够提高多媒体资源的推荐效果。

Description

多媒体资源的推荐方法及装置
交叉引用
本申请主张2017年8月3日提交的中国专利申请号为201710656161.X的优先权，其全部内容通过引用包含于此。
技术领域
本公开涉及多媒体技术领域，尤其涉及一种多媒体资源的推荐方法及装置。
背景技术
在互联网时代，尤其是移动互联网时代，如何为用户提供及时且有价值的信息是众多互联网公司研究的热点。近年来，随着机器学习系统的发展，推荐系统开始支持个性化推荐策略。个性化推荐策略需要结合不同用户的实际使用情况进行差异化处理。
如何向用户推荐用户更感兴趣的多媒体资源，以提高多媒体资源的推荐效果，是亟待解决的问题。
发明内容
有鉴于此，本公开提出了一种多媒体资源的推荐方法及装置，以提高多媒体资源的推荐效果。
根据本公开的一方面，提供了一种多媒体资源的推荐方法，包括：
确定多个待选多媒体资源；
确定各个所述待选多媒体资源的点击量预估值；
根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；
根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。
在一种可能的实现方式中，确定待选多媒体资源，包括：
根据所述目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，确定待选多媒体资源，包括：
根据所述目标用户所属用户集群中各个用户的用户行为数据，确定所述用户集群对应的可选多媒体资源；
将所述可选多媒体资源中所述目标用户未观看的多媒体资源确定为待选多媒体资 源。
在一种可能的实现方式中，确定待选多媒体资源，包括：
根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，确定各个所述待选多媒体资源的点击量预估值，包括：
将所述待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源；
对于每个所述第一类待选多媒体资源，根据所述第一类待选多媒体资源的历史点击数据确定所述待选第一类多媒体资源的点击量预估值；
对于每个所述第二类待选多媒体资源，查找与所述第二类待选多媒体资源匹配的第一类待选多媒体资源，计算所述第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据所述匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算所述第二类待选多媒体资源的点击量预估值。
在一种可能的实现方式中，根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源，包括：
获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；
根据各个推荐窗口对应的推荐场景，以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定针对各个推荐窗口的待推荐的多媒体资源。
在一种可能的实现方式中，根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐，包括：
根据所述待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从所述待推荐的多媒体资源中确定推荐结果；
将所述推荐结果推荐给所述目标用户。
根据本公开的另一方面，提供了一种多媒体资源的推荐装置，包括：
第一确定模块，用于确定多个待选多媒体资源；
第二确定模块，用于确定各个所述待选多媒体资源的点击量预估值；
第三确定模块，根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；
推荐模块，用于根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。
在一种可能的实现方式中，所述第一确定模块包括：
第一确定子模块，用于根据所述目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，所述第一确定模块包括：
第二确定子模块，用于根据所述目标用户所属用户集群中各个用户的用户行为数据，确定所述用户集群对应的可选多媒体资源；
第三确定子模块，用于将所述可选多媒体资源中所述目标用户未观看的多媒体资源确定为待选多媒体资源。
在一种可能的实现方式中，所述第一确定模块包括：
第四确定子模块，用于根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，所述第二确定模块包括：
划分子模块，用于将所述待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源；
第五确定子模块，用于对于每个所述第一类待选多媒体资源，根据所述第一类待选多媒体资源的历史点击数据确定所述待选第一类多媒体资源的点击量预估值；
第六确定子模块，用于对于每个所述第二类待选多媒体资源，查找与所述第二类待选多媒体资源匹配的第一类待选多媒体资源，计算所述第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据所述匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算所述第二类待选多媒体资源的点击量预估值。
在一种可能的实现方式中，所述第三确定模块包括：
获取子模块，用于获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；
第七确定子模块，用于根据各个推荐窗口对应的推荐场景，以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定针对各个推荐窗口的待推荐的多媒体资源。
在一种可能的实现方式中，所述推荐模块包括：
第八确定子模块，用于根据所述待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从所述待推荐的多媒体资源中确定推荐结果；
推荐子模块，用于将所述推荐结果推荐给所述目标用户。
根据本公开的另一方面，提供了一种多媒体资源的推荐装置，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为执行上述方法。
根据本公开的另一方面，提供了一种非易失性计算机可读存储介质，其上存储有计算机程序指令，其中，所述计算机程序指令被处理器执行时实现上述方法。
本公开的各方面的多媒体资源的推荐方法及装置通过确定多个待选多媒体资源，确定各个待选多媒体资源的点击量预估值，根据推荐场景以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定待推荐的多媒体资源，并根据待推荐的多媒体 资源向目标用户进行多媒体资源推荐，由此能够提高多媒体资源的推荐效果。
根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
包含在说明书中并且构成说明书的一部分的附图与说明书一起示出了本公开的示例性实施例、特征和方面，并且用于解释本公开的原理。
图1示出根据本公开一实施例的多媒体资源的推荐方法的流程图。
图2示出根据本公开一实施例的多媒体资源的推荐方法步骤S12的一示例性的流程图。
图3示出根据本公开一实施例的多媒体资源的推荐方法步骤S121的一示例性的流程图。
图4示出根据本公开一实施例的多媒体资源的推荐方法步骤S121的另一示例性的流程图。
图5示出根据本公开一实施例的多媒体资源的推荐方法步骤S122的一示例性的流程图。
图6示出根据本公开一实施例的多媒体资源的推荐方法步骤S123的一示例性的流程图。
图7示出根据本公开一实施例的多媒体资源的推荐装置的框图。
图8示出根据本公开一实施例的多媒体资源的推荐装置的一示例性的框图。
图9是根据一示例性实施例示出的一种用于多媒体资源的推荐的装置1900的框图。
具体实施方式
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
另外，为了更好的说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。
图1示出根据本公开一实施例的多媒体资源的推荐方法的流程图。该方法可以应用于服务器中，在此不作限定。如图1所示，该方法包括步骤S11至步骤S14。
在步骤S11中，确定多个待选多媒体资源。
在本实施例中，多媒体可以为多种媒体的综合，例如包括文本、声音和图像。多媒体资源可以为视频和音频等中的一项或多项，在此不作限定。
在一种可能的实现方式中，确定待选多媒体资源，可以包括：根据目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从多媒体资源库的各个多媒体资源中确定待选多媒体资源。
作为该实现方式的一个示例，目标用户的标签可以根据目标用户的用户行为数据确定。目标用户的用户行为数据可以包括但不限于以下至少一项：目标用户观看的多媒体资源、目标用户评论的多媒体资源、目标用户发表弹幕的多媒体资源、目标用户购买的多媒体资源、目标用户订阅的多媒体资源、目标用户收藏的多媒体资源和目标用户点赞的多媒体资源。例如，可以根据目标用户的用户行为数据中的各个多媒体资源的标题和简介中的一项或两项，确定目标用户的标签。
作为该实现方式的一个示例，多媒体资源库可以指某一视频网站中存储所有或者部分多媒体资源的数据库。
作为该实现方式的一个示例，多媒体资源的标签可以根据多媒体资源的标题和简介中的一项或两项确定。例如，可以对多媒体资源的标题和简介进行分词，根据分词结果确定多媒体资源的标签。
作为该实现方式的一个示例，可以计算目标用户的标签与各个多媒体资源的标签的相似度，并可以根据目标用户的标签与各个多媒体资源的标签的相似度确定待选多媒体资源。例如，可以将与目标用户的标签的相似度大于第一阈值的多媒体资源确定为待选多媒体资源。又如，可以将与目标用户的标签的相似度最大的N个多媒体资源确定为待选多媒体资源，其中，N为正整数。
作为该实现方式的一个示例，可以基于内容推荐模型来实现该实现方式。
该实现方式根据目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从多媒体资源库的各个多媒体资源中确定待选多媒体资源，由此确定的待选多媒体资源能够反映目标用户的真实兴趣，从而有助于提高多媒体资源推荐的准确性和效果。
在另一种可能的实现方式中，确定待选多媒体资源，可以包括：根据目标用户所属用户集群中各个用户的用户行为数据，确定该用户集群对应的可选多媒体资源；将可选多媒体资源中目标用户未观看的多媒体资源确定为待选多媒体资源。例如，该用户集群 对应的可选多媒体资源为C1、C2、C3、C4、C5和C6，其中，目标用户已观看的多媒体资源为C3，则可以将C1、C2、C4、C5和C6确定为待选多媒体资源。
作为该实现方式的一个示例，可以将用户行为数据相似的多个用户归于同一个用户集群中。其中，若多个用户的用户行为数据相似，则可以表示该多个用户观看多媒体资源的喜好相似。
作为该实现方式的另一个示例，可以将标签相同或相似的多个用户归于同一个用户集群中。
作为该实现方式的另一个示例，可以根据目标用户的请求，将目标用户加入目标用户感兴趣的用户集群中。
作为该实现方式的一个示例，根据目标用户所属用户集群中各个用户的用户行为数据，确定该用户集群对应的可选多媒体资源，可以包括：将该用户集群中各个用户的用户行为数据中的多媒体资源确定为该用户集群对应的可选多媒体资源。
作为该实现方式的一个示例，可以基于多媒体资源的协同过滤模型来实现该实现方式。
该实现方式根据目标用户所属用户集群中各个用户的用户行为数据，确定该用户集群对应的可选多媒体资源，并将可选多媒体资源中目标用户未观看的多媒体资源确定为待选多媒体资源，由此能够通过用户集群的用户行为数据挖掘目标用户可能感兴趣的多媒体资源，从而有助于提高多媒体资源推荐的准确性和效果。
在另一种可能的实现方式中，确定待选多媒体资源，可以包括：根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从多媒体资源库的各个多媒体资源中确定待选多媒体资源。
作为该实现方式的一个示例，可以根据多媒体资源库中的各个多媒体资源在指定时间段内的搜索量、点击量和点击率中的一项或多项，确定各个多媒体资源的热点得分，并将热点得分最高的M个多媒体资源确定为待选多媒体资源，其中，M为正整数。例如，指定时间段可以为3天内。
例如，可以采用式3计算多媒体资源库中的某一多媒体资源在指定时间段内的热点得分S r：
S r＝α 1×R 1+α 2×R 2+α 3×R 3  式3；
其中，R 1表示该多媒体资源在指定时间段内的搜索量，α 1表示R 1对应的权重；R 2表示该多媒体资源在指定时间段内的点击量，α 2表示R 2对应的权重；R 3表示该多媒体资源 在指定时间段内的点击率，α 3表示R 3对应的权重。
作为该实现方式的一个示例，可以分别将多个多媒体资源分类中热点得分最高的P个多媒体资源确定为待选多媒体资源，其中，P为正整数。例如，可以将电影分类中热点得分最高的P个多媒体资源、电视剧分类中热点得分最高的P个多媒体资源、综艺分类中热点得分最高的P个多媒体资源和体育分类中热点得分最高的P个多媒体资源分别确定为待选多媒体资源。
作为该实现方式的一个示例，可以基于热点模型来实现该实现方式。
该实现方式根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从多媒体资源库的各个多媒体资源中确定待选多媒体资源，由此能够将热点的多媒体资源确定为待选多媒体资源，从而有助于提高多媒体资源推荐的效果。
在步骤S12中，确定各个待选多媒体资源的点击量预估值。
在步骤S13中，根据推荐场景以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定待推荐的多媒体资源。
作为本实施例的一个示例，可以结合待选多媒体资源的点击量预估值和推荐场景，计算各个待选多媒体资源的得分，并可以将得分最高的L个待选多媒体资源确定为待推荐的多媒体资源。其中，推荐场景可以包括热门推荐、个性化推荐、频道推荐和节日相关推荐等中的一种或多种。在推荐场景为热门推荐的情况下，可以根据待选多媒体资源的实时搜索量，以及待选多媒体资源的点击量预估值，计算各个待选多媒体资源的得分。在推荐场景为个性化推荐的情况下，可以根据待推荐的多媒体资源的标签与目标用户的标签的匹配度，以及待选多媒体资源的点击量预估值，计算各个待选多媒体资源的得分。在推荐场景为频道推荐的情况下，可以根据待选多媒体资源与频道的匹配度，以及待选多媒体资源的点击量预估值，计算各个待选多媒体资源的得分。其中，频道可以包括电视剧频道、新闻频道、综艺频道和体育频道等。同一个待选多媒体资源在不同频道中的得分可能不同。在推荐场景为节日推荐的情况下，可以根据待推荐的多媒体资源与当前节日的匹配度，以及待选多媒体资源的点击量预估值，计算各个待选多媒体资源的得分。例如，在春节期间，可以向目标用户推荐与过春节相关的多媒体资源。
在一种可能的实现方式中，还可以根据应用平台来调整进行推荐的待选多媒体资源的个数。其中，应用平台可以包括网页端和App(Application，应用)端。
在一种可能的实现方式中，根据推荐场景以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定待推荐的多媒体资源，包括：获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；根据各个推荐窗口对应的推荐场景，以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定针对各个推荐窗口的待推荐的 多媒体资源。例如，推荐窗口包括第一推荐窗口、第二推荐窗口、第三推荐窗口和第四推荐窗口。其中，第一推荐窗口对应的推荐场景为热门推荐，第二推荐窗口对应的推荐场景为个性化推荐，第三推荐窗口对应的推荐场景为节日相关推荐，第四推荐窗口对应的推荐场景为电视剧频道推荐。
在步骤S14中，根据待推荐的多媒体资源向目标用户进行多媒体资源推荐。
在一种可能的实现方式中，可以在接收到来自于目标用户的推荐请求的情况下，根据待推荐的多媒体资源向目标用户进行多媒体资源推荐。
作为本实施例的一个示例，可以将得分最高的一个待推荐的多媒体资源推荐给目标用户。
作为本实施例的另一个示例，可以将得分最高的多个待推荐的多媒体资源推荐给目标用户。
作为本实施例的另一个示例，可以将所有待推荐的多媒体资源推荐给目标用户。
在一种可能的实现方式中，根据待推荐的多媒体资源向目标用户进行多媒体资源推荐，可以包括：根据待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从待推荐的多媒体资源中确定推荐结果；将所述推荐结果推荐给所述目标用户。
作为该实现方式的一个示例，推荐结果中可以最多包括P个同一上传者上传的多媒体资源，以提高推荐结果的多样性。其中，P为正整数。例如，P等于3。
作为该实现方式的另一个示例，推荐结果中可以最多包括Q个同一二级频道的多媒体资源，以提高推荐结果的多样性。其中，Q为正整数。例如，Q等于3。例如，综艺频道为某一一级频道，湖南综艺频道为该一级频道下的二级频道。
作为该实现方式的另一个示例，推荐结果中可以最多包括R个具有同一标签的多媒体资源，以提高推荐结果的多样性。其中，R为正整数。例如，R等于3。
本实施例通过确定多个待选多媒体资源，确定各个待选多媒体资源的点击量预估值，根据推荐场景以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定待推荐的多媒体资源，并根据待推荐的多媒体资源向目标用户进行多媒体资源推荐，由此能够提高多媒体资源的推荐效果。
图2示出根据本公开一实施例的多媒体资源的推荐方法步骤S12的一示例性的流程图。如图2所示，步骤S12可以包括步骤S121至S123。
在步骤S121中，将待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源。
在本实施例中，可以将所有待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源，并对于第一类待选多媒体资源和第二类待选多媒体资源分别采用不同 的方法计算其点击量预估值。
在步骤S122中，对于每个第一类待选多媒体资源，根据第一类待选多媒体资源的历史点击数据确定待选第一类多媒体资源的点击量预估值。
在步骤S123中，对于每个第二类待选多媒体资源，查找与第二类待选多媒体资源匹配的第一类待选多媒体资源，计算第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算第二类待选多媒体资源的点击量预估值。
例如，对于第二类待选多媒体资源v 1，若查找到与该第二类待选多媒体资源v 1匹配的第一类待选多媒体资源v 2，则计算第二类待选多媒体资源v 1与第一类待选多媒体资源v 2的匹配度，再根据匹配度以及第一类待选多媒体资源v 2的点击量预估值计算第二类待选多媒体资源v 1的点击量预估值。
图3示出根据本公开一实施例的多媒体资源的推荐方法步骤S121的一示例性的流程图。如图3所示，步骤S121可以包括步骤S1211和S1212。
在步骤S1211中，分别获取每个待选多媒体资源的历史点击数据。
在步骤S1212中，对于每个待选多媒体资源，根据待选多媒体资源的历史点击数据判断待选多媒体资源的当前点击量是否处于下降期，若是，则将待选多媒体资源确定为第一类待选多媒体资源，否则将待选多媒体资源确定为第二类待选多媒体资源。
其中，待选多媒体资源的当前点击量处于下降期可以指该待选多媒体资源随着时间的推后其点击量呈现下降的趋势。
图4示出根据本公开一实施例的多媒体资源的推荐方法步骤S121的另一示例性的流程图。如图4所示，步骤S121可以包括步骤S1213和S1214。
在步骤S1213中，分别获取每个待选多媒体资源的上传时间。
在步骤S1214中，对于每个待选多媒体资源，判断待选多媒体资源的上传时间距离当前系统时间的时间长度是否大于第二阈值，若是，则将待选多媒体资源确定为第一类待选多媒体资源，否则将待选多媒体资源确定为第二类待选多媒体资源。
例如，第二阈值可以为48小时，在此不作限定。
图5示出根据本公开一实施例的多媒体资源的推荐方法步骤S122的一示例性的流程图。如图5所示，步骤S122可以包括步骤S1221至S1223。
在步骤S1221中，对于每个第一类待选多媒体资源，获取第一类待选多媒体资源的历史点击数据。
在步骤S1222中，根据第一类待选多媒体资源的历史点击数据训练第一类待选多媒体资源的时间衰减系数。
在步骤S1223中，根据第一类待选多媒体资源的历史点击数据和时间衰减系数计算第一类待选多媒体资源的点击量预估值。
在一种可能的实现方式中，根据该第一类待选多媒体资源的历史点击数据训练该第一类待选多媒体资源的时间衰减系数，可以包括：根据第一类待选多媒体资源i的历史点击数据，采用式1训练第一类待选多媒体资源i的时间衰减系数θ，其中，第一类待选多媒体资源i的历史点击数据包括第一类待选多媒体资源i在指定日期之前的点击量；
f(t)＝V i0×e θt  式1；
其中，V i0表示第一类待选多媒体资源i在指定日期的点击量；t＝0表示指定日期，f(0)＝V i0；当t＞0时，t表示在指定日期前的第t天，f(t)表示第一类待选多媒体资源i在指定日期前的第t天的点击量；当t＜0时，t表示在指定日期后的第-t天，f(t)表示第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值。
例如，第一类待选多媒体资源i的历史点击数据包括第一类待选多媒体资源i在指定日期的点击量以及第一类待选多媒体资源i在指定日期之前15天的点击量。t＝1表示指定日期的前一天，f(1)表示第一类待选多媒体资源i在指定日期的前一天的点击量，t＝2表示指定日期的前两天，f(2)表示第一类待选多媒体资源i在指定日期的前两天的点击量，以此类推，t＝15表示指定日期的前15天，f(15)表示第一类待选多媒体资源i在指定日期的前15天的点击量。在采用式1训练第一类待选多媒体资源i的时间衰减系数θ时，对于t＞0，式1中的f(t)、V i0和t是已知的，根据f(t)、V i0和t可以训练得到第一类待选多媒体资源i的时间衰减系数θ。根据第一类待选多媒体资源i的历史点击数据还可以对时间衰减系数θ进行校正。
在该实现方式中，根据该第一类待选多媒体资源的历史点击数据和时间衰减系数计算该第一类待选多媒体资源的点击量预估值，可以包括：采用训练后的式1计算第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值。
t＝-1表示指定日期的前一天，f(-1)表示第一类待选多媒体资源i在指定日期的前一天的点击量。在训练得到第一类待选多媒体资源i的时间衰减系数θ后，对于t＜0，式1中的V i0、θ和t是已知的，从而能计算出在指定日期后的某天的点击量预估值。例如， 采用训练后的式1计算第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值，可以为：采用训练后的式1计算f(-1)，将f(-1)确定为第一类待选多媒体资源i的点击量预估值。
在另一种可能的实现方式中，根据该第一类待选多媒体资源的历史点击数据训练该第一类待选多媒体资源的时间衰减系数，可以包括：根据第一类待选多媒体资源i的历史点击数据，采用式2训练第一类待选多媒体资源i的时间衰减系数θ，其中，第一类待选多媒体资源i的历史点击数据包括第一类待选多媒体资源i在指定日期之前的点击量；
g(t)＝lg(V i0×e θt)  式2；
其中，V i0表示第一类待选多媒体资源i在指定日期的点击量；t＝0表示指定日期，g(0)＝lgV i0；当t＞0时，t表示在指定日期前的第t天，g(t)表示第一类待选多媒体资源i在指定日期前的第t天的点击量的对数值；当t＜0时，t表示在指定日期后的第-t天，g(t)表示第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值的对数值。
在该实现方式中，根据该第一类待选多媒体资源的历史点击数据和时间衰减系数计算该第一类待选多媒体资源的点击量预估值，包括：采用训练后的式2计算第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值的对数值。
在该实现方式中，对点击量取对数，从而降低数值，便于计算和存储。在该示例中，根据式2计算得到的第一类待选多媒体资源i在指定日期后的第-t天的点击量预估值的对数值可以作为第一类待选多媒体资源i的rank(排序)值。通过计算新上传的多媒体资源的rank值，能够让更多新上传的优质的多媒体资源得到曝光。
图6示出根据本公开一实施例的多媒体资源的推荐方法步骤S123的一示例性的流程图。如图6所示，步骤S123可以包括步骤S1231和S1232。
在步骤S1231中，对于每个第二类待选多媒体资源，根据第二类待选多媒体资源的指定信息查找与第二类待选多媒体资源匹配的第一类待选多媒体资源；指定信息包括以下至少一项：上传时间、时间长度和上传者信息。
在步骤S1232中，根据第二类待选多媒体资源的指定信息和所匹配的第一类待选多媒体资源的指定信息计算第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度。
其中，第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度大于0且小于或等于1。
作为本实施例的一个示例，第二类待选多媒体资源与第一类待选多媒体资源的上传 时间越接近，则两者之间的匹配度越高。
作为本实施例的另一个示例，第二类待选多媒体资源与第一类待选多媒体资源的时间长度越接近，则两者之间的匹配度越高。例如，若该待选多媒体资源为视频，则该待选多媒体资源的时间长度可以指的是视频长度。
作为本实施例的另一个示例，若第二类待选多媒体资源与第一类待选多媒体资源的上传者相同，则第二类待选多媒体资源与第一类待选多媒体资源的匹配度较高。
作为本实施例的另一个示例，第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度可以根据上传时间、时间长度和上传者信息确定，其中，上传时间对于匹配度的权重为λ 1，时间长度对于匹配度的权重为λ 2，上传者信息对于匹配度的权重为λ 3。
在一种可能的实现方式中，根据匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算该第二类待选多媒体资源的点击量预估值，可以包括：将匹配度与所匹配的第一类待选多媒体资源的点击量预估值的乘积作为该第二类待选多媒体资源的点击量预估值。
图7示出根据本公开一实施例的多媒体资源的推荐装置的框图。如图7所示，该装置包括：第一确定模块71，用于确定多个待选多媒体资源；第二确定模块72，用于确定各个所述待选多媒体资源的点击量预估值；第三确定模块73，根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；推荐模块74，用于根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。
图8示出根据本公开一实施例的多媒体资源的推荐装置的一示例性的框图。如图8所示：
在一种可能的实现方式中，所述第一确定模块71包括：第一确定子模块711，用于根据所述目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，所述第一确定模块71包括：第二确定子模块712，用于根据所述目标用户所属用户集群中各个用户的用户行为数据，确定所述用户集群对应的可选多媒体资源；第三确定子模块713，用于将所述可选多媒体资源中所述目标用户未观看的多媒体资源确定为待选多媒体资源。
在一种可能的实现方式中，所述第一确定模块71包括：第四确定子模块714，用于根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
在一种可能的实现方式中，所述第二确定模块72包括：划分子模块721，用于将所述 待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源；第五确定子模块722，用于对于每个所述第一类待选多媒体资源，根据所述第一类待选多媒体资源的历史点击数据确定所述待选第一类多媒体资源的点击量预估值；第六确定子模块723，用于对于每个所述第二类待选多媒体资源，查找与所述第二类待选多媒体资源匹配的第一类待选多媒体资源，计算所述第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据所述匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算所述第二类待选多媒体资源的点击量预估值。
在一种可能的实现方式中，第三确定模块73包括：获取子模块731，用于获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；第七确定子模块732，用于根据各个推荐窗口对应的推荐场景，以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定针对各个推荐窗口的待推荐的多媒体资源。
在一种可能的实现方式中，所述推荐模块74包括：第八确定子模块741，用于根据所述待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从所述待推荐的多媒体资源中确定推荐结果；推荐子模块742，用于将所述推荐结果推荐给所述目标用户。
本实施例通过确定多个待选多媒体资源，确定各个待选多媒体资源的点击量预估值，根据推荐场景以及各个待选多媒体资源的点击量预估值，从各个待选多媒体资源中确定待推荐的多媒体资源，并根据待推荐的多媒体资源向目标用户进行多媒体资源推荐，由此能够提高多媒体资源的推荐效果。
图9是根据一示例性实施例示出的一种用于多媒体资源的推荐的装置1900的框图。例如，装置1900可以被提供为一服务器。参照图9，装置1900包括处理组件1922，其进一步包括一个或多个处理器，以及由存储器1932所代表的存储器资源，用于存储可由处理组件1922的执行的指令，例如应用程序。存储器1932中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的模块。此外，处理组件1922被配置为执行指令，以执行上述方法。
装置1900还可以包括一个电源组件1926被配置为执行装置1900的电源管理，一个有线或无线网络接口1950被配置为将装置1900连接到网络，和一个输入输出(I/O)接口1958。装置1900可以操作基于存储在存储器1932的操作系统，例如Windows ServerTM，Mac OS XTM，UnixTM,LinuxTM，FreeBSDTM或类似。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1932，上述计算机程序指令可由装置1900的处理组件1922执行以完成上述方法。
本公开可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本公开的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或 框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。

Claims (16)

  1. 一种多媒体资源的推荐方法，其特征在于，包括：
    确定多个待选多媒体资源；
    确定各个所述待选多媒体资源的点击量预估值；
    根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；
    根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。
  2. 根据权利要求1所述的方法，其特征在于，确定待选多媒体资源，包括：
    根据所述目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
  3. 根据权利要求1所述的方法，其特征在于，确定待选多媒体资源，包括：
    根据所述目标用户所属用户集群中各个用户的用户行为数据，确定所述用户集群对应的可选多媒体资源；
    将所述可选多媒体资源中所述目标用户未观看的多媒体资源确定为待选多媒体资源。
  4. 根据权利要求1所述的方法，其特征在于，确定待选多媒体资源，包括：
    根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
  5. 根据权利要求1所述的方法，其特征在于，确定各个所述待选多媒体资源的点击量预估值，包括：
    将所述待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源；
    对于每个所述第一类待选多媒体资源，根据所述第一类待选多媒体资源的历史点击数据确定所述待选第一类多媒体资源的点击量预估值；
    对于每个所述第二类待选多媒体资源，查找与所述第二类待选多媒体资源匹配的第一类待选多媒体资源，计算所述第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据所述匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算所述第二类待选多媒体资源的点击量预估值。
  6. 根据权利要求1所述的方法，其特征在于，根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源，包括：
    获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；
    根据各个推荐窗口对应的推荐场景，以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定针对各个推荐窗口的待推荐的多媒体资源。
  7. 根据权利要求1所述的方法，其特征在于，根据所述待推荐的多媒体资源向目标 用户进行多媒体资源推荐，包括：
    根据所述待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从所述待推荐的多媒体资源中确定推荐结果；
    将所述推荐结果推荐给所述目标用户。
  8. 一种多媒体资源的推荐装置，其特征在于，包括：
    第一确定模块，用于确定多个待选多媒体资源；
    第二确定模块，用于确定各个所述待选多媒体资源的点击量预估值；
    第三确定模块，根据推荐场景以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定待推荐的多媒体资源；
    推荐模块，用于根据所述待推荐的多媒体资源向目标用户进行多媒体资源推荐。
  9. 根据权利要求8所述的装置，其特征在于，所述第一确定模块包括：
    第一确定子模块，用于根据所述目标用户的标签，以及多媒体资源库中各个多媒体资源的标签，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
  10. 根据权利要求8所述的装置，其特征在于，所述第一确定模块包括：
    第二确定子模块，用于根据所述目标用户所属用户集群中各个用户的用户行为数据，确定所述用户集群对应的可选多媒体资源；
    第三确定子模块，用于将所述可选多媒体资源中所述目标用户未观看的多媒体资源确定为待选多媒体资源。
  11. 根据权利要求8所述的装置，其特征在于，所述第一确定模块包括：
    第四确定子模块，用于根据多媒体资源库中各个多媒体资源的搜索量、点击量和点击率中的一项或多项，从所述多媒体资源库的各个多媒体资源中确定待选多媒体资源。
  12. 根据权利要求8所述的装置，其特征在于，所述第二确定模块包括：
    划分子模块，用于将所述待选多媒体资源划分为第一类待选多媒体资源和第二类待选多媒体资源；
    第五确定子模块，用于对于每个所述第一类待选多媒体资源，根据所述第一类待选多媒体资源的历史点击数据确定所述待选第一类多媒体资源的点击量预估值；
    第六确定子模块，用于对于每个所述第二类待选多媒体资源，查找与所述第二类待选多媒体资源匹配的第一类待选多媒体资源，计算所述第二类待选多媒体资源与所匹配的第一类待选多媒体资源的匹配度，并根据所述匹配度与所匹配的第一类待选多媒体资源的点击量预估值计算所述第二类待选多媒体资源的点击量预估值。
  13. 根据权利要求8所述的装置，其特征在于，所述第三确定模块包括：
    获取子模块，用于获取多媒体资源的推荐页面中各个推荐窗口对应的推荐场景；
    第七确定子模块，用于根据各个推荐窗口对应的推荐场景，以及各个所述待选多媒体资源的点击量预估值，从各个所述待选多媒体资源中确定针对各个推荐窗口的待推荐的多媒体资源。
  14. 根据权利要求8所述的装置，其特征在于，所述推荐模块包括：
    第八确定子模块，用于根据所述待推荐的多媒体资源的上传者、所属的频道和标签中的至少一项，从所述待推荐的多媒体资源中确定推荐结果；
    推荐子模块，用于将所述推荐结果推荐给所述目标用户。
  15. 一种多媒体资源的推荐装置，其特征在于，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器被配置为执行权利要求1至7中任意一项所述的方法。
  16. 一种非易失性计算机可读存储介质，其上存储有计算机程序指令，其特征在于，所述计算机程序指令被处理器执行时实现权利要求1至7中任意一项所述的方法。
PCT/CN2018/095586 2017-08-03 2018-07-13 多媒体资源的推荐方法及装置 WO2019024670A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710656161.X 2017-08-03
CN201710656161.XA CN109388739A (zh) 2017-08-03 2017-08-03 多媒体资源的推荐方法及装置

Publications (1)

Publication Number Publication Date
WO2019024670A1 true WO2019024670A1 (zh) 2019-02-07

Family

ID=65232722

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/095586 WO2019024670A1 (zh) 2017-08-03 2018-07-13 多媒体资源的推荐方法及装置

Country Status (2)

Country Link
CN (1) CN109388739A (zh)
WO (1) WO2019024670A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114201626A (zh) * 2021-11-18 2022-03-18 北京达佳互联信息技术有限公司 多媒体推荐方法、装置、电子设备及存储介质

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110598045B (zh) * 2019-09-06 2021-03-19 腾讯科技（深圳）有限公司 视频推荐方法、装置
CN110889020B (zh) * 2019-11-22 2022-08-23 百度在线网络技术（北京）有限公司 站点资源挖掘方法、装置以及电子设备
CN114547429A (zh) * 2020-11-23 2022-05-27 北京达佳互联信息技术有限公司 数据推荐方法、装置、服务器及存储介质
CN112667906A (zh) * 2020-12-31 2021-04-16 上海众源网络有限公司 一种up主的推荐方法、装置及电子设备
CN114780828A (zh) * 2022-02-28 2022-07-22 北京达佳互联信息技术有限公司 资源推荐方法、装置、计算机设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104333773A (zh) * 2013-12-18 2015-02-04 乐视网信息技术（北京）股份有限公司 一种视频推荐方法及服务器
CN104504059A (zh) * 2014-12-22 2015-04-08 合一网络技术(北京)有限公司 多媒体资源推荐方法
CN105956086A (zh) * 2016-04-29 2016-09-21 合网络技术（北京）有限公司 多媒体资源的推荐方法和装置
CN106227834A (zh) * 2016-07-26 2016-12-14 合网络技术（北京）有限公司 多媒体资源的推荐方法及装置
CN106294830A (zh) * 2016-08-17 2017-01-04 合智能科技（深圳）有限公司 多媒体资源的推荐方法及装置

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4124115B2 (ja) * 2003-12-02 2008-07-23 ソニー株式会社 情報処理装置及び情報処理方法、並びにコンピュータ・プログラム
CN102938861A (zh) * 2012-09-28 2013-02-20 深圳市龙视传媒有限公司 一种交互式数字电视门户展示装置、方法及终端
CN104572688A (zh) * 2013-10-17 2015-04-29 腾讯科技（深圳）有限公司 信息推送方法及装置
CN103648046A (zh) * 2013-12-20 2014-03-19 乐视致新电子科技（天津）有限公司 智能电视中的视频推荐页面展示方法及系统
CN103731738A (zh) * 2014-01-23 2014-04-16 哈尔滨理工大学 基于用户群组行为分析的视频推荐方法及装置
CN105989004B (zh) * 2015-01-27 2020-04-14 阿里巴巴集团控股有限公司 一种信息投放的预处理方法和装置
CN105005582B (zh) * 2015-06-17 2017-09-15 深圳市腾讯计算机系统有限公司 多媒体信息的推荐方法及装置
CN105095431A (zh) * 2015-07-22 2015-11-25 百度在线网络技术（北京）有限公司 根据用户的行为信息推送视频的方法和装置
CN105392065A (zh) * 2015-12-28 2016-03-09 Tcl集团股份有限公司 一种智能电视页面内容布局的方法及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104333773A (zh) * 2013-12-18 2015-02-04 乐视网信息技术（北京）股份有限公司 一种视频推荐方法及服务器
CN104504059A (zh) * 2014-12-22 2015-04-08 合一网络技术(北京)有限公司 多媒体资源推荐方法
CN105956086A (zh) * 2016-04-29 2016-09-21 合网络技术（北京）有限公司 多媒体资源的推荐方法和装置
CN106227834A (zh) * 2016-07-26 2016-12-14 合网络技术（北京）有限公司 多媒体资源的推荐方法及装置
CN106294830A (zh) * 2016-08-17 2017-01-04 合智能科技（深圳）有限公司 多媒体资源的推荐方法及装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114201626A (zh) * 2021-11-18 2022-03-18 北京达佳互联信息技术有限公司 多媒体推荐方法、装置、电子设备及存储介质
CN114201626B (zh) * 2021-11-18 2023-03-28 北京达佳互联信息技术有限公司 多媒体推荐方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN109388739A (zh) 2019-02-26

Similar Documents

Publication Publication Date Title
WO2019024670A1 (zh) 多媒体资源的推荐方法及装置
US10795939B2 (en) Query method and apparatus
CN108694223B (zh) 一种用户画像库的构建方法及装置
US10579675B2 (en) Content-based video recommendation
US9594826B2 (en) Co-selected image classification
US10305851B1 (en) Network-based content discovery using messages of a messaging platform
CN111178970B (zh) 广告投放的方法及装置、电子设备和计算机可读存储介质
US10747771B2 (en) Method and apparatus for determining hot event
US20160203193A1 (en) Context aware query selection
CN109582862B (zh) 点击率预估方法、介质、系统和计算设备
US20170195753A1 (en) System and method for generating segmented content based on related data ranking
JP2018530847A (ja) 広告配信のための動画使用情報処理
CN112100489B (zh) 对象推荐的方法、装置和计算机存储介质
EP4092545A1 (en) Content recommendation method and device
CN109255037B (zh) 用于输出信息的方法和装置
CN108462900B (zh) 视频推荐方法及装置
US11921732B2 (en) Artificial intelligence and/or machine learning systems and methods for evaluating audiences in an embedding space based on keywords
WO2018192272A1 (zh) 多媒体资源的推荐方法及装置
US11341138B2 (en) Method and system for query performance prediction
WO2018214493A1 (zh) 视频搜索方法及装置
CN113705683B (zh) 推荐模型的训练方法、装置、电子设备及存储介质
CN115858815A (zh) 确定映射信息的方法、广告推荐方法、装置、设备及介质
CN114580790A (zh) 生命周期阶段预测和模型训练方法、装置、介质及设备
CN113360761A (zh) 信息流推荐方法、装置、电子设备和计算机可读存储介质
CN109657129B (zh) 用于获取信息的方法及装置

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18840622

Country of ref document: EP

Kind code of ref document: A1