WO2017101299A1 - 视频推荐方法、装置和设备 - Google Patents

视频推荐方法、装置和设备 Download PDF

Info

Publication number
WO2017101299A1
WO2017101299A1 PCT/CN2016/088113 CN2016088113W WO2017101299A1 WO 2017101299 A1 WO2017101299 A1 WO 2017101299A1 CN 2016088113 W CN2016088113 W CN 2016088113W WO 2017101299 A1 WO2017101299 A1 WO 2017101299A1
Authority
WO
WIPO (PCT)
Prior art keywords
video
category
user
preference value
user identifier
Prior art date
Application number
PCT/CN2016/088113
Other languages
English (en)
French (fr)
Inventor
关涛
Original Assignee
乐视控股（北京）有限公司
乐视网信息技术（北京）股份有限公司
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 乐视控股（北京）有限公司, 乐视网信息技术（北京）股份有限公司 filed Critical 乐视控股（北京）有限公司
Priority to US15/247,758 priority Critical patent/US20170169040A1/en
Publication of WO2017101299A1 publication Critical patent/WO2017101299A1/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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/73—Querying
    • G06F16/735—Filtering based on additional data, e.g. user or group profiles

Definitions

  • the present application belongs to the field of Internet, and in particular, to a video recommendation method, apparatus and device.
  • the current home page content displayed after the video application is opened is the home page data set by the server.
  • the home page includes various categories, such as TV, movie, and animation. , original, etc., when the user wants to watch a certain video program, it is usually necessary to input the video name or the performer's name as a keyword in the search bar, and then find the desired video from the search result.
  • the video application also has a video similar to the "user favorite" on the homepage to recommend videos to the user, but the content is small and is likely to be a video that the user has already watched.
  • the page displaying the video in a waterfall layout manner Usually there is no special recommended area, so it is more uncomfortable Use existing recommendations.
  • the embodiment of the present application provides a video recommendation method, apparatus, and device, which are used to solve the technical problem that the video application in the prior art is less effective in recommending a video to a user.
  • the present application discloses a video recommendation method, the method includes: classifying videos, and sorting videos in each category according to popularity of the video; analyzing each user identifier according to the browsing history a preference value of each category; obtaining a preference value of the logged-in user identifier for each category according to a user identifier of the logged-in terminal device and a preference value of each user identifier for each category; The video of each category is extracted for each category's preference level value and the sorted result and pushed to the terminal device for presentation.
  • the present application further discloses a video recommendation apparatus, the apparatus comprising: a video classification module, configured to classify videos, and sort videos in each category according to popularity of the video; An analysis module, configured to analyze, according to the browsing record, a preference value of each user identifier for each category; the data obtaining module is configured to obtain, according to the user identifier of the login terminal device and the preference value of each user identifier for each category, The logged-in user identifies a preference value for each category; the video push module is configured to pull the video under each category according to the preference value of each category and the sorted result according to the logged-in user identifier, and push the video to the The terminal device is displayed.
  • a video classification module configured to classify videos, and sort videos in each category according to popularity of the video
  • An analysis module configured to analyze, according to the browsing record, a preference value of each user identifier for each category
  • the data obtaining module is configured to obtain, according to the user identifier of the login terminal device and the preference value
  • the present application further discloses a video recommendation device, including: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to: classify a video, and according to The popularity of the video ranks the videos in each category; the preference value of each user identifier for each category is analyzed according to the browsing history; and the preference value of each category according to the user identifier of the login terminal device and the respective user identifiers, Obtaining a preference value of the user ID of the login for each category; extracting a video under each category according to the preference value of each category and the result of the sorting, and pushing the video to the terminal The device is displayed.
  • the server sorts and sorts the video and analyzes the user preference, determines the user's preference according to the registered user identifier, pushes the corresponding category, and ranks.
  • the top video to the terminal device displays the pushed video in the video application homepage, so that the user can directly see the video that matches his or her preference. Since the pushed video is the top ranked video in each category, the push video is The popularity can also be improved, and the user's manual input of keywords for searching is eliminated, which is more convenient for users.
  • FIG. 1 is a flowchart of a video recommendation method according to an embodiment of the present application
  • FIG. 2 is a block diagram of a video recommendation apparatus according to an embodiment of the present application.
  • the background server classifies the video and sorts the videos in each category, and analyzes the degree of preference of each user for each category. After the user logs in to the server, the user is determined according to the registered user identifier. The degree is combined with the video sorting results under each category to push the corresponding number of videos to the terminal device for display. After the background video classification and user analysis, the effect of personalized recommendation to the user is improved, and the recommended video is more suitable for the user. Likes.
  • FIG. 1 is a video recommendation method provided by an embodiment of the present application, which is applicable to a server, and the server may be a background server corresponding to a video application. As shown in FIG. 1, the method includes the following steps S10-S13.
  • step S10 the videos are classified, and the videos under the respective categories are sorted according to the popularity of the videos.
  • the video can be classified according to the feature information such as the video name, the performer and the like, the image features of the video image, and the like.
  • the same video can also be grouped into different categories at the same time. For example, for an entertainment video, it may be classified into the “Entertainment” and “Information” categories at the same time.
  • Popularity can be related to factors such as the number of clicks on a video, ratings, and number of comments. Through the above factors, the popularity of the video is comprehensively determined, and the videos in each category are sorted according to the popularity, and the videos ranked first can be preferentially recommended to the user.
  • step S11 the preference value of each user identifier for each category is analyzed based on the browsing history.
  • the browsing record is a video viewing record saved by the server corresponding to each user ID (UserID), including the video that appears in the user's page and the video that the user clicked on. In combination with the classification of the video in step S10, it is analyzed which category the video in the browsing record belongs to, so that the user is interested in which category of video.
  • UserID user ID
  • the stored videos have been classified, and combined with the analysis of the user's browsing records, the preference value of each user identifier for each category is obtained.
  • its preference value for all categories can be collectively referred to as the user's user portrait.
  • each class is based on the user identifier of the login terminal device and each user identifier.
  • the other preference value is obtained as the value of the preference of the registered user identifier for each category.
  • the terminal device After the terminal device opens the video application, the user enters a user identifier, a password, and a verification code to complete the login.
  • the server obtains the preference value of the user identifier for each category according to the registered user identifier.
  • step S13 the video under each category is pulled and pushed to the terminal device for display according to the value of the preference of each category and the result of the ranking according to the registered user identifier.
  • the size of the preference value determines the probability of pulling the video from the corresponding category, and in combination with the result of the sorting in step S10, the video with the highest ranking is preferentially pushed to the video application of the terminal device for display.
  • the video application's home page you can display 20 video entries, the user logo "Zhang San", the preference value for the entertainment video is 0.4, and the fitness rating is 0.5, then the video displayed on the home page. It includes 8 entertainment videos, and 10 sports videos, and is the video of the top 8 in the entertainment category and the top 10 in the sports category in the sorting result of step S10, respectively.
  • the terminal device can be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a car console, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and the like.
  • the pushed video is displayed in the home page of the video application installed on the terminal device, and the pushed video can be displayed by the waterfall flow layout.
  • the visual representation of the waterfall flow layout is a jagged multi-column layout that, as the page scroll bar scrolls down, will continue to load a preset number of video items and append to the current tail.
  • the server continues to pull the video to the terminal device according to the above rules.
  • the server sorts and sorts the video and analyzes the user's preference, determines the user's preference according to the registered user identifier, and pushes the corresponding category and the top ranked video to the terminal device, and displays the push in the video application homepage.
  • Video so that users can directly see the video that suits their preferences. Since the pushed video is the top-ranked video in each category, the popularity of the push video can be improved, and the user can manually remove the key. Word search operations are more user-friendly.
  • the popularity of the video may be reflected by a composite score, step S10 Further embodiments may be as follows in the following steps S101-S103.
  • step S101 feature information of the video is acquired.
  • the feature information may be text features such as a video title, a video source, and a video content introduction, for example, a character name, a team name, a place name, a building name, a game name, and the like appearing in the video title, and a television station or website in the video source.
  • a character name, team name, place name, building name, game name and other information in the video content introduction can be used as text features for video classification.
  • the feature information may also be image features from a video image, such as image features of sports games, animations, news, movies, etc., identified using image recognition techniques; or audio features derived from audio recognition, such as recognized jazz Audio features such as music, pop music, symphony, drama, and cross talk.
  • step S102 the video is classified according to the feature information by using a preset classification algorithm.
  • the preset classification algorithm may be a classification algorithm that performs matching according to the feature information, or a classifier trained for a different category through a specific training set, for example, a Support Vector Machine (SVM).
  • SVM Support Vector Machine
  • step S103 the composite score for each video is calculated under each category, and sorted according to the composite score from high to low.
  • the composite score can be calculated by the following formula:
  • BaseScore(video) Hotness(video) ⁇ Fresshness(video), where BaseScore(video) represents the composite score of the video, Hotness(video) represents the heat of the video, and Freshness(video) represents the newness of the video.
  • the heat that is, the more times the video is clicked and viewed in a certain period of time, the higher the heat; the newness, that is, the closer the video is released to the current time, the higher the newness, the possibility that the user has not seen the video.
  • the values of heat and freshness can be from 1.0 to 10.0 to one digit after the decimal point.
  • the sorting result under the “entertainment” category is video C, video B, and video A. Considering the heat and the newness, the video C will be pushed to the user first.
  • the comprehensive scores of the videos are calculated and ranked by the heat and the newness pairs, and the number of viewers in each category and the newly released videos are preferentially pushed to the user, and the possibility that the users have not watched the videos is very High, so it is more attractive to users who like the corresponding category, and the fit of the pushed video and user interest is further improved.
  • step S11 may be further implemented as the following steps S111-S113.
  • step S111 the exposure amount and the click amount corresponding to the video of each category are obtained.
  • the exposure refers to the number of video entries of a certain category displayed on the user page
  • the click volume refers to the number of times the video item of a certain category displayed is opened by the user to watch the corresponding video.
  • the user ID "ABC123” has displayed a total of 500 "entertainment” categories of videos in its page, and the number of these 500 videos selected by the user "ABC123” is 150 (including the same video being viewed multiple times).
  • the number of times, then, the video of the "Entertainment” category corresponds to the user "ABC123” with an exposure of 500 and a click volume of 150.
  • the exposure and the amount of clicks of the user identification for each category are determined in the above manner.
  • step S112 the preference value of the user identifier for the category is determined according to the ratio of the click amount to the exposure amount, that is,
  • the category indicates the category
  • user indicates the user
  • Click indicates the user user's click on the video under the category
  • Exposure indicates the exposure of the video under the category category when the user is logged in.
  • the video of the "entertainment” category corresponds to the user "ABC123" with an exposure amount of 500 and the click amount is 150. Then, substituting the above formula, the user "ABC123" has a preference for the "entertainment” category of 30%.
  • the user's identification value for each category can be calculated by the above formula.
  • step S113 the preference value of the category is normalized, namely:
  • NormaizeFavorite (user, category) represents the normalized preference value
  • MaxFavorite (user, category) represents the maximum preference value of the user user for each category.
  • the purpose of the normalization process is to reflect the push probability of new videos appearing in the corresponding category, and to ensure that new videos under the user's favorite category can be pushed.
  • the user "ABC123” has a preference for the "entertainment” category of 30%, a preference for the "soccer” category of 60%, and a preference for the "news” category of 45%, then the preference value is the largest.
  • the category is "soccer”.
  • the user can obtain the category most interested in the user and set the video push probability of this category to 1. That is, the newly released video of this category will definitely be pushed to the user, which improves the push efficiency of the video of the user's favorite category.
  • the push probability of each category will also be adjusted.
  • the video of the category with the highest degree of value has the highest probability of pushing.
  • step S13 may be further exemplified by the following steps S131-S133.
  • step S131 the video is pulled down from each category according to the value of the preference of each category and the result of the ranking according to the registered user identifier.
  • the priority value of each category is determined, and the probability that the preference value or the normalized preference value is larger is the probability of being pulled. The higher.
  • the video sorting results of each category the corresponding videos are sequentially pulled until a preset number (for example, 30) of videos is pulled.
  • step S132 the video that has been exposed to the user identification is filtered out in the pulled video.
  • the captured video is exposed (appeared) in the user's page within a preset duration (for example, within one week) according to the browsing record corresponding to the user identifier, or is received by the user within the preset duration Watched. If the captured video has been exposed or viewed, it indicates that the video has been pushed to the user recently. The user has already seen the video. At this time, the corresponding video is filtered from the captured video. According to the sorting result, the video of the corresponding category continues to pull other videos for replacement until the above conditions are met, so that the video that has not been pushed recently is pushed to the user, and the video that the user has recently seen is prevented from being lowered in the pushed video. The user's desire to watch.
  • a preset duration for example, within one week
  • step S133 the filtered video is pushed to the terminal device for presentation.
  • the captured video is filtered, and the video that has been pushed in the preset duration is filtered from the captured video, so that the pushed video is a video that the user has not seen recently to enhance the push.
  • the viewing rate of the video is a video that the user has not seen recently to enhance the push.
  • step S13 may be further implemented as the following steps S134-S136.
  • step S134 the video is pulled down from each category according to the value of the preference of each category and the result of the sorting according to the registered user identifier.
  • step S135 the captured video is scored, and the number of videos consecutively belonging to the same category in the sorting result is less than or equal to the preset number.
  • the value can be between 1.0-10.0, and for negative factors, the value can be between 0.01-0.99.
  • Positive factors may include heat, timeliness, views, etc.
  • negative factors may include junk index, pornographic index, number of reports, and the like.
  • the junk index indicates that the content of the video is not welcomed by netizens or that the video image or sound is flawed, and the viewing effect may not be satisfactory;
  • the pornographic index represents the degree to which the video is not viewed in adulthood;
  • the number of reports represents that the video has bad or illegal information. Net The number of times the friend reported.
  • Score(user,video) BaseScore(video) ⁇ UserFavorite(category,video) ⁇ Freshness(vidoe) ⁇ ...
  • BaseScore(video) is the composite score of the video
  • UserFavorite(category,video) is the user Freshness (video) is the newness of the video
  • the last ellipsis can also consider other positive and negative factors together to score the captured video.
  • the captured video is scored in the above manner, and sorted according to the score result from high to low, and pushed to the terminal device in the order of the sort result.
  • a threshold value may be set for the number of consecutive similar videos, for example, four, when the number of consecutive similar videos is greater than four, in the first Before the five similar videos, insert one or more other top-scoring videos to score the diversity, thus achieving diversity adjustment and preventing the category of push video from being too single.
  • the top five videos are sports categories, including video 1 to video 5, the sixth is the entertainment category video 6, and the seventh is the news category.
  • Video 7, ranked 8th is the sports category video 8.
  • the order of the video 6 whose ranking result is higher is adjusted to be between the video 4 and the video 5, and the adjusted sort result is as follows: video 1 , Video 2, Video 3, Video 4, Video 6, Video 5, Video 7, Video 8.
  • step S136 the sorted video is pushed to the terminal device for presentation.
  • the captured video is scored and sorted according to the scored result, and the higher scored video is preferentially recommended to the user, so that the better quality video program is preferentially pushed, and the diversity adjustment can be performed in the sorting process.
  • the higher scored video is preferentially recommended to the user, so that the better quality video program is preferentially pushed, and the diversity adjustment can be performed in the sorting process.
  • other categories of videos are inserted therein to maintain the diversity of push videos.
  • the above method of filtering the exposed video may be combined with the method of sorting and sorting the video, and filtering the video that the user has recently seen in the captured video, and simultaneously Maintain the diversity of push videos.
  • FIG. 2 is a block diagram of a video recommendation apparatus according to an embodiment of the present disclosure.
  • the device is located on the server side, and includes a video classification module 20, a user analysis module 21, a data acquisition module 22, and a video push module 23.
  • the video classification module 20 is electrically connected to the user analysis module 21 for classifying the video, and sorting the videos in each category according to the popularity of the video;
  • the user analysis module 21 is electrically connected to the data acquisition module 22, and is configured to analyze, according to the browsing record, a preference value of each user identifier for each category;
  • the data acquisition module 22 is electrically connected to the video push module 23, and is configured to obtain, according to the user identifier of the login terminal device and the preference value of each user identifier for each category, the preference value of the registered user identifier for each category;
  • the video pushing module 23 is configured to pull the video under each category according to the value of the preference of each category and the result of the sorting according to the registered user identifier, and push the video to the terminal device for display.
  • the popularity of the video is a composite score of the video
  • the video classification module 20 further includes: a first acquisition sub-module, a classification sub-module, and a first sequencing sub-module.
  • the first obtaining submodule is electrically connected to the classifying submodule, and is configured to acquire feature information of the video;
  • the classification sub-module is electrically connected to the first sorting sub-module, and is configured to classify the video according to the characteristic information by using a preset classification algorithm;
  • the first sorting sub-module is configured to calculate a composite score for each video under each category, and sort according to the composite score from high to low.
  • the first sorting submodule includes:
  • BaseScore(video) Hotness(video) ⁇ Freshness(video); where BaseScore(video) represents the composite score of the video, Hotness(video) represents the heat of the video, and Freshness(video) represents the newness of the video.
  • the user analysis module 21 further includes: a second acquisition sub-module, a determination sub-module and a normalization sub-module.
  • the second obtaining sub-module is electrically connected to the determining sub-module, and is configured to obtain an exposure amount and a click amount corresponding to the video and the user identifier in each category;
  • the determining submodule is electrically connected to the normalized submodule, and is configured to determine a preference value of the user identifier for the category according to a ratio of the click amount to the exposure amount, that is,
  • the category indicates the category
  • the user indicates the user
  • the click (user, category) indicates the user user's click on the video in the category category
  • the Exposure (user, category) indicates the exposure of the video under the category category when the user user logs in;
  • a normalization sub-module for normalizing the preference value of the category namely:
  • NormaizeFavorite (user, category) represents the normalized preference value
  • MaxFavorite (user, category) represents the maximum preference value of the user user for each category.
  • the video push module 23 further includes: a first pull submodule, a filter submodule, and a first push submodule.
  • the first pull submodule is electrically connected to the filter submodule, and is configured to pull a video from each category according to the preference value of the category and the result of the sorting according to the registered user identifier;
  • the filtering sub-module is electrically connected to the first pushing sub-module, and is configured to filter out the video that has been exposed to the user identifier in the captured video;
  • the first push sub-module is configured to push the filtered video to the terminal device for display.
  • the video push module 23 further includes: a second pull submodule, a second sort submodule, and a second push submodule.
  • the second pull submodule is electrically connected to the second sorting submodule, and is configured to pull a video from each category according to the preference value of the category and the result of the sorting according to the registered user identifier;
  • the second sorting sub-module is electrically connected to the second push sub-module, and is used for sorting and sorting the captured video, and the number of consecutive videos belonging to the same category in the sorting result is less than or equal to a preset number;
  • the second push sub-module is configured to push the sorted video to the terminal device for display.
  • each of the foregoing functional modules may be implemented by a hardware processor.
  • the embodiment of the present application further provides a video recommendation device, where the device includes: a processor and a memory for storing processor executable instructions;
  • the processor is configured to: classify the video, and sort the videos in each category according to the popularity of the video; analyze the preference value of each user identifier for each category according to the browsing record; according to the login terminal device a user identifier and a preference value of each user identifier for each category, obtaining a preference value of the registered user identifier for each category; and extracting each category according to the preference level value and the sorted result
  • the video is pushed to the terminal device for display.
  • the popularity of the video is a comprehensive score of the video
  • the sorting the video, and sorting the videos in each category according to the comprehensive score of the video comprises: acquiring feature information of the video; using a preset classification algorithm And classifying the video according to the feature information; calculating a composite score of each video under each category, and sorting according to the comprehensive score from high to low.
  • the calculation of the composite score for each video under each category includes:
  • BaseScore(video) Hotness(video) ⁇ Freshness(video); where BaseScore(video) represents the composite score of the video, Hotness(video) represents the heat of the video, and Freshness(video) represents the newness of the video.
  • the analyzing, according to the browsing record, the preference value of each user identifier for each category includes: obtaining an exposure amount and a click amount corresponding to the video of each category according to the user identifier; determining, according to a ratio of the click amount to the exposure amount, The value of the user's preference for the category, ie:
  • category indicates the category
  • user indicates the user
  • Click indicates the user user's click on the video in the category category
  • Exposure indicates the category category. The amount of exposure of the video under the user user login
  • NormaizeFavorite (user, category) represents the normalized preference value
  • MaxFavorite (user, category) represents the maximum preference value of the user user for each category.
  • the extracting the video under each category according to the preference level value and the sorted result and pushing the video to the terminal device for displaying includes: a preference degree value of each category according to the registered user identifier and a result of the sorting
  • the video is pulled down from each category; the video that has been exposed to the user identifier is filtered out in the captured video; and the filtered video is pushed to the terminal device for display.
  • the extracting the video under each category according to the preference level value and the sorted result and pushing the video to the terminal device for displaying includes: a preference degree value of each category according to the registered user identifier and a result of the sorting Pulling the video from each category; sorting and sorting the captured video, and the number of consecutive videos belonging to the same category in the sorting result is less than or equal to the preset number; and pushing the sorted video to the terminal device for display.
  • the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, ie may be located A place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. Those of ordinary skill in the art can understand and implement without deliberate labor.
  • the method according to the present application can also be implemented as a computer program executed by a CPU, which can be stored in a computer readable storage medium.
  • the computer program is executed by the CPU, the above-described functions defined in the method of the present application are performed.
  • the method steps and system units described above may also be implemented with a controller and a computer readable storage medium for storing a computer program that causes the controller to implement the steps or unit functions described above.
  • non-volatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash.
  • ROM read only memory
  • PROM programmable ROM
  • EPROM electrically programmable ROM
  • EEPROM electrically erasable programmable ROM
  • flash flash.
  • Volatile memory can include random access memory (RAM), which can act as external cache memory.
  • RAM can be obtained in a variety of forms, such as synchronous RAM (DRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM) and direct Rambus RAM (DRRAM).
  • DRAM synchronous RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDR SDRAM double data rate SDRAM
  • ESDRAM enhanced SDRAM
  • SLDRAM Synchronous Link DRAM
  • DRRAM direct Rambus RAM
  • Storage devices of the disclosed aspects are intended to comprise, without being limited to, these and other suitable types of memory.
  • DSPs digital signal processors
  • ASIC dedicated An integrated circuit
  • FPGA field programmable gate array
  • a general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine.
  • the processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.
  • a software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
  • An exemplary storage medium is coupled to the processor such that the processor can be Read information or write information to the storage medium.
  • the storage medium can be integrated with a processor.
  • the processor and the storage medium can reside in an ASIC.
  • the ASIC can reside in the user terminal.
  • the processor and the storage medium may reside as a separate component in the server.
  • the functions may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored on or transmitted as one or more instructions or code on a computer readable medium.
  • Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one location to another.
  • a storage medium may be any available media that can be accessed by a general purpose or special purpose computer.
  • the computer readable medium may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage device, disk storage device or other magnetic storage device, or may be used to carry or store a form of instructions Or the required program code of the data structure and any other medium that can be accessed by a general purpose or special purpose computer or a general purpose or special purpose processor. Also, any connection is properly termed a computer-readable medium.
  • a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technology such as infrared, radio, and microwave is used to transmit software from a website, server, or other remote source
  • the coaxial line Cables, fiber optic cables, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are all included in the definition of the medium.
  • a magnetic disk and an optical disk include a compact disk (CD), a laser disk, an optical disk, a digital versatile disk (DVD), a floppy disk, a Blu-ray disk, in which a disk generally reproduces data magnetically, and the optical disk optically reproduces data using a laser. . Combinations of the above should also be included within the scope of computer readable media.
  • a person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium.
  • the storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.
  • the video is sorted and sorted, and the user's preference is analyzed, the user's preference is determined according to the registered user identifier, and the video of the corresponding category and ranked top is pushed to the terminal device.
  • the video on the home page of the video application shows the pushed video, so that users can directly see the video that suits their preferences. Since the pushed video is the top-ranked video in each category, the popularity of the push video can be improved, and It eliminates the user's manual input of keywords for searching, which is more convenient for users to use.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Multimedia (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)

Abstract

一种视频推荐方法、装置和设备，该方法包括：对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序（S10）；根据浏览记录分析各个用户标识对各个类别的喜好程度值（S11）；根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值（S12）；根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至终端设备进行展示（S13）。使推送视频的受欢迎程度能够得到提升，而且免去了用户手动输入关键词进行搜索的操作，更加便于用户使用。

Description

视频推荐方法、装置和设备
交叉引用
本申请引用于2015年12月15日递交的名称为“视频推荐方法、装置和服务器”的第201510938060.2号中国专利申请，其通过引用被全部并入本申请。
技术领域
本申请属于互联网领域，具体地说，涉及一种视频推荐方法、装置和设备。
背景技术
随着智能手机和平板电脑等移动设备的高速增长，带动了移动视频应用的发展，同时也带来了移动数据流量的激增。目前移动视频已经成为移动互联网的主力应用之一，占据了所有移动数据流量的59％，移动视频已成为移动流量增长的驱动力。研究表明，观看电视的用户数和用户在电视上消耗的时间都在逐年递减，而同时按需观看视频和通过移动终端观看视频的用户数正在呈爆发式增长。
随着移动视频用户数量的激增，移动视频应用市场的竞争也越发激烈，用户对于移动视频应用的要求也越来越高。经过研究表明，用户打开视频应用主要是以浏览视频为主，目前的视频应用程序打开后所显示的主页内容是服务器设置好的主页数据，主页中包括各种分类，例如，电视，电影，动漫，原创等等，用户想要观看某个视频节目时，通常需要将视频名称或者表演者姓名做为关键词输入在搜索栏中，再从搜索结果中查找到想看的视频。
目前视频应用程序虽然也有在主页设置一个类似于“用户喜欢”的区域向用户推荐视频，但内容较少并且很可能是用户已经看过的视频，在以瀑布流布局方式展示视频的页面中，通常没有设置专门的推荐区域，因此更不适 合使用现有的推荐方式。
发明内容
有鉴于此，本申请实施例提供了一种视频推荐方法、装置和设备，用以解决现有技术中视频应用程序向用户个性化推荐视频的效果较差的技术问题。
为了解决上述技术问题，本申请公开了一种视频推荐方法，所述方法包括：对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；根据浏览记录分析各个用户标识对各个类别的喜好程度值；根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示。
为了解决上述技术问题，本申请还公开了一种视频推荐装置，所述装置包括：视频分类模块，用于对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；用户分析模块，用于根据浏览记录分析各个用户标识对各个类别的喜好程度值；数据获取模块，用于根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；视频推送模块，用于根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示。
为了解决上述技术问题，本申请还公开了一种视频推荐设备，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：对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；根据浏览记录分析各个用户标识对各个类别的喜好程度值；根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端 设备进行展示。
与现有技术相比，本申请实施例提供的视频推荐方法、装置和设备，服务器对视频进行分类排序并对用户喜好进行分析，根据登录的用户标识确定用户的喜好，推送相应类别的且排名靠前的视频至终端设备，在视频应用程序首页中展示推送的视频，从而使用户能够直接看到符合自己喜好的视频，由于推送的视频是各个类别中排序靠前的视频，因此推送视频的受欢迎程度也能够得到提升，而且免去了用户手动输入关键词进行搜索的操作，更加便于用户使用。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本申请。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述中的附图是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例提供的一种视频推荐方法的流程图；
图2是本申请实施例提供的一种视频推荐装置的框图。
通过上述附图，已示出本申请明确的实施例，后文中将有更详细的描述。这些附图和文字描述并不是为了通过任何方式限制本申请构思的范围，而是通过参考特定实施例为本领域技术人员说明本申请的概念。
具体实施方式
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获 得的所有其他实施例，都属于本申请保护的范围。
本申请实施例，后台服务器对视频进行分类并对各类别下的视频进行排序，分析每个用户对各类别的喜好程度，在用户登录服务器之后，根据登录的用户标识确定其对各类别的喜好程度并结合各类别下的视频排序结果推送相应数量的视频至终端设备进行展示，经过后台的视频分类和用户分析，提高了向用户进行视频个性化推荐的效果，推荐的视频更贴合用户的喜好。
图1是本申请实施例提供的一种视频推荐方法，适用于服务器，服务器可以是与视频应用程序对应的后台服务器。如图1所示，该方法包括以下步骤S10-S13。
在步骤S10中，对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序。
可根据视频名称、表演者等文字特征，视频图像的图像特征等特征信息对视频进行分类。同一个视频也可以同时被分到不同的类别中，例如，对于一条娱乐资讯视频来说，就可能被同时分到“娱乐”以及“资讯”类别中。
受欢迎程度可以与视频的点击数、评分、评论数等因素相关。通过以上因素来综合确定出视频的受欢迎程度，根据受欢迎程度对各个类别下的视频进行排序，排序靠前的视频能够被优先推荐给用户。
在步骤S11中，根据浏览记录分析各个用户标识对各个类别的喜好程度值。
浏览记录是服务器保存的与每个用户标识(UserID)对应的视频观看记录，包括在用户的页面中出现过的视频和用户点击观看过的视频。结合步骤S10中对视频的分类分析出浏览记录中的视频分别属于哪一类别，从而得到该用户对哪一类别的视频比较感兴趣。
此时，在后台服务器中，已对存储的视频进行了分类，并结合分析用户的浏览记录得到了每个用户标识分别对各个类别的喜好程度值。对一个用户标识而言，其对所有类别的喜好程度值可共同被称为该用户的用户画像。
在步骤S12中，根据登录终端设备的用户标识和各个用户标识对各个类 别的喜好程度值，获取登录的用户标识对各个类别的喜好程度值。
用户在终端设备打开视频应用程序后，输入用户标识、密码以及验证码等信息完成登录，服务器根据登录的用户标识获取该用户标识对各个类别的喜好程度值。
在步骤S13中，根据登录的用户标识对各个类别的喜好程度值和排序的结果拉取各个分类下的视频并推送至终端设备进行展示。
喜好程度值的大小决定了从相应类别拉取视频的几率的高低，结合步骤S10中排序的结果，将排序靠前的视频优先推送至终端设备的视频应用程序进行展示。
例如，在视频应用程序的首页可展示20个视频条目，用户标识“张三”，对娱乐类的视频的喜好程度值为0.4，对体育类的喜好程度值为0.5，那么在首页展示的视频包括8个娱乐类视频，以及10个体育类视频，并且分别是在步骤S10的排序结果中在娱乐类别排序前8的视频和在体育类别中排序前10的视频。
终端设备可以是移动电话，计算机，数字广播终端，消息收发设备，车载控制台，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等。在终端设备安装的视频应用程序的首页中展示推送的视频，可以通过瀑布流布局的方式来展示推送的视频。瀑布流布局的视觉表现为参差不齐的多栏布局，随着页面滚动条向下滚动，这种布局还会不断加载预设数量的视频条目并附加至当前尾部。当再次加载数据时，服务器继续按照上述规则拉取视频推送至终端设备。
本实施例中，服务器对视频进行分类排序并对用户喜好进行分析，根据登录的用户标识确定用户的喜好，推送相应类别的且排名靠前的视频至终端设备，在视频应用程序首页中展示推送的视频，从而使用户能够直接看到符合自己喜好的视频，由于推送的视频是各个类别中排序靠前的视频，因此推送视频的受欢迎程度也能够得到提升，而且免去了用户手动输入关键词进行搜索的操作，更加便于用户使用。
在一个实施例中，可通过综合分数来反映视频的受欢迎程度，步骤S10 可进一步被实施例为以下步骤S101-S103。
在步骤S101中，获取所述视频的特征信息。
特征信息可以是来自视频标题、视频来源、视频内容介绍等文字特征，例如，视频标题中的出现的人物姓名、球队名称、地名、建筑名称、比赛名称等信息，视频来源中的电视台、网站等信息，视频内容介绍中的人物姓名、球队名称、地名、建筑名称、比赛名称等信息都可以做为文字特征用于视频分类。
特征信息也可以是来自视频图像的图像特征，例如利用图像识别技术识别出的体育比赛、动画、新闻、电影等图像特征；还可以是根据音频识别而得到的音频特征，例如，识别出的爵士音乐、流行音乐、交响乐、戏剧、相声等音频特征。
在步骤S102中，利用预设的分类算法并根据特征信息对视频进行分类。
预设的分类算法可以是根据特征信息进行匹配的分类算法，也可以是针对不同类别通过特定的训练集而训练出来的分类器，例如，支持向量机(Support Vector Machine，SVM)。
在步骤S103中，在各个类别下计算每个视频的综合分数，根据综合分数由高到低进行排序。
综合分数可以通过以下公式计算：
BaseScore(video)＝Hotness(video)×Fresshness(video)，其中，BaseScore(video)代表视频的综合分数，Hotness(video)代表视频的热度，Freshness(video)代表视频的时新性。
热度，即视频在一定时间内被点击观看的次数越多，其热度越高；时新性，即视频的发布时间距当前时间越近，时新性越高，用户没有看过该视频的可能性也就越高。
热度和时新性的取值可以是从1.0至10.0的精确到小数点后一位的数值。例如，“娱乐”类别下的视频A、视频B和视频C。视频A的热度Hotness(A)＝6.0，时新性Freshness(A)＝4.0，那么，视频A的综合分数BaseScore (A)＝24。视频B的热度Hotness(B)＝7.5，时新性Freshness(A)＝6.8，那么，视频B的综合分数BaseScore(B)＝51。视频C的热度Hotness(C)＝8.8，时新性Freshness(C)＝7，那么，视频C的综合分数BaseScore(C)＝61.6。那么“娱乐”类别下的的排序结果即为视频C、视频B、视频A，综合考虑热度和时新性，视频C将会被优先推送给用户。
本实施例中，通过热度和时新性对来计算视频的综合分数并排序，可以将各个类别中观看人数多并且是新发布的视频优先推送给用户，用户没有观看过这些视频的可能性很高，因此对于喜好相应类别的用户而言具有较高的吸引力，推送的视频与用户兴趣的贴合度得到进一步提升。
在一个实施例中，步骤S11可进一步实施为以下步骤S111-S113。
在步骤S111中，获取各个类别下的视频与用户标识对应的曝光量和点击量。
曝光量是指用户页面中展示的某个类别的视频条目的个数，点击量是指所展示的某个类别的视频条目中被该用户打开观看相应的视频的次数。例如，用户标识“ABC123”，在其页面中共展示过500个“娱乐”类别的视频，而这500个视频中被用户“ABC123”选择观看的次数为150次(包括同一个视频被多次观看的次数)，那么，“娱乐”类别的视频与用户“ABC123”对应的曝光量为500，点击量为150。
通过以上方式确定用户标识针对各个类别的曝光量和点击量。
在步骤S112中，根据点击量与曝光量的比值确定用户标识对所述类别的喜好程度值，即
Figure PCTCN2016088113-appb-000001
其中category表示类别，user表示用户，Click(user,category)表示用户user对category类别下的视频的点击量，Exposure(user,category)表示category类别下的视频在用户user登录时的曝光量。
例如，上例中“娱乐”类别的视频与用户“ABC123”对应的曝光量为500，点击量为150，那么，代入上述公式，用户“ABC123”对“娱乐”类别的喜好程度为30％。
用户标识针对各个类别的喜好程度值都可以通过上述公式计算得出。
在步骤S113中，对所述类别的喜好程度值进行归一化处理，即：
Figure PCTCN2016088113-appb-000002
其中，NormaizeFavorite(user,category)代表归一化喜好程度值，MaxFavorite(user,category)代表用户user对各个类别category的最大喜好程度值。
进行归一化处理的目的是为了反映出相应的类别下出现新视频的推送几率，保证用户最喜欢的类别下的新视频能够被推送。
例如，用户“ABC123”对“娱乐”类别的喜好程度为30％，对“足球”类别的喜好程度为60％，对“新闻”类别的喜好程度值为45％，那么，喜好程度值最大的类别为“足球”，根据上述公式，用户“ABC123”对“娱乐”类别的归一化喜好程度值为30％/60％＝0.5，对“足球”类别的归一化喜好程度值为60％/60％＝1，对“新闻”类别的归一化喜好程度值为45％/60％＝0.75。因此，如果“足球”类别下发布了新的视频，则肯定会推送给该用户，而“娱乐”类别下发布了新的视频，则会有0.5的可能性推送给该用户，“新闻”类别下发布了新的视频，则会有0.75的可能性推送给该用户。
本实施例中，通过用户对各个类别视频的曝光量和点击量计算出喜好程度值并进行归一化处理，能够得到该用户最感兴趣的类别并将这一类别的视频推送几率置为1，即该类别新发布的视频肯定会推送给该用户，提升了对用户最喜好类别的视频的推送效率，当然随着用户浏览行为的变化，各个类别的推送几率也会随之调整，对于喜好程度值最高的类别的视频，推送的几率也相应最高。
在一个实施例中，步骤S13可进一步被实施例为以下步骤S131-S133。
在步骤S131中，根据登录的用户标识对各个类别的喜好程度值和排序的结果从各个分类下拉取视频。
首先根据登录的用户标识对各个类别的喜好程度值确定优先拉取的类别，喜好程度值或者归一化喜好程度值越大的类别下的视频，被拉取的几率 越高。然后，再按照各个类别的视频排序结果依次拉取相应的视频，直到拉取预设数量(例如30个)的视频。
在步骤S132中，在拉取的视频中过滤掉已对用户标识曝光过的视频。
根据用户标识对应的浏览记录查询拉取到的视频是否在预设时长内(例如，一周之内)在用户的页面中曝光过(出现过)相应的条目，或者在该预设时长内被用户观看过。如果拉取的视频被曝光过或者被观看过，说明该视频在最近已经向用户推送过，用户已经见到过该视频，此时，将相应的视频从拉取到的视频中过滤掉，在相应类别的视频中按照排序结果继续拉取其他视频进行替换，直到符合上述条件，从而将最近没有被推送过的视频推送给用户，防止推送的视频中存在用户最近看到过的视频而降低了用户的观看欲望。
在步骤S133中，将过滤后的视频推送至终端设备进行展示。
本实施例中，对拉取的视频进行过滤，将在预设时长内推送过的视频从拉取的视频中过滤掉，使推送的视频都是用户最近没有见到过的视频，以提升推送的视频的观看率。
在一个实施例中，步骤S13还可以进一步实施为以下步骤S134-S136。
在步骤S134中，根据登录的用户标识对各个类别的喜好程度值和排序的结果从各个分类下拉取视频。
在步骤S135中，对拉取的视频进行打分排序，并且排序结果中连续属于同一类别的视频数小于或等于预设数量。
对拉取到的视频进行打分时需要考虑多种因素，包括能提高打分分数的正面因子和降低打分分数的负面因子。对于正面因子而言，其取值可以在1.0-10.0之间，对于负面因子而言，其取值可以在0.01-0.99之间。
正面因子可包括热度、时效性、观看次数等，而负面因子可包括垃圾指数、色情指数、举报次数等。垃圾指数代表该视频的内容不受网友欢迎或者是视频图像或者声音有瑕疵，观看效果可能不理想；色情指数代表该视频不事宜未成年观看的程度；举报次数代表该视频因存在不良或非法信息而被网 友举报的次数。
例如，对拉取到的视频进行打分可使用以下公式：
Score(user,video)＝BaseScore(video)×UserFavorite(category,video)×Freshness(vidoe)×...其中，BaseScore(video)为视频的综合分数，UserFavorite(category,video)是该用户对该视频所属类别的喜好程度值，Freshness(video)是该视频的时新性，最后省略号代表还可以一起综合考虑其他正面因子和负面因子，为拉取到的视频打分。例如，视频A的综合分数为24，该用户对相应的类别的喜好程度值为50％，视频的时效性为4.0，热度为6.0，垃圾指数和色情指数均为0.95，那么该视频打分结果Score＝24×50％×4.0×6.0×0.95×0.95＝259.92。
通过上述方式对拉取到的视频进行打分，并根据打分结果由高至低进行排序，按照排序结果的顺序推送至终端设备。在排序结果中，有可能连续出现同一类别下的视频，此时，可以对连续出现同类视频的数量设置一门限值，例如4个，当连续出现同类视频的数量大于4个时，在第五个同类视频之前，插入一个或多个其他打分结果排序最靠前的视频，从而实现多样性调整，防止推送视频的类别过于单一。
例如，在打分的排序结果中，排在前五位的视频都是体育类别，包括视频1至视频5，排在第6位的是娱乐类别的视频6，排在第七位的是新闻类别的视频7，排在第8位的是体育类别的视频8。此时，由于连续出现的体育类别的视频数量已经超过了4个，因此要将排序结果更靠前的视频6的顺序调整到视频4和视频5之间，调整后的排序结果如下：视频1、视频2、视频3、视频4、视频6、视频5、视频7、视频8。
在步骤S136中，将排序后的视频推送至终端设备进行展示。
本实施例中，对拉取的视频进行打分并根据打分结果排序，将打分更高的视频优先推荐给用户，从而将更优质的视频节目优先进行推送，在排序过程中还可以进行多样性调整，在相同类别的视频连续出现的次数过多时，在其中插入其他类别的视频，以保持推送视频的多样性。
此外，还可以将上述对已曝光过的视频进行过滤的方法和对视频打分排序并进行多样性调整的方法相结合，在拉取的视频中过滤掉用户最近看到过的视频，同时，还保持了推送视频的多样性。
下述为本申请装置实施例，可以用于执行本申请方法实施例。
图2是本申请实施例提供的一种视频推荐装置的框图，位于服务器侧，该装置包括视频分类模块20，用户分析模块21，数据获取模块22和视频推送模块23。
视频分类模块20与用户分析模块21电连接，用于对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；
用户分析模块21与数据获取模块22电连接，用于根据浏览记录分析各个用户标识对各个类别的喜好程度值；
数据获取模块22与视频推送模块23电连接，用于根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取登录的用户标识对各个类别的喜好程度值；
视频推送模块23，用于根据登录的用户标识对各个类别的喜好程度值和排序的结果拉取各个分类下的视频并推送至终端设备进行展示。
在一个实施例中，所述视频的受欢迎程度为视频的综合分数，该视频分类模块20进一步包括：第一获取子模块，分类子模块和第一排序子模块。
第一获取子模块与分类子模块电连接，用于获取视频的特征信息；
分类子模块与第一排序子模块电连接，用于利用预设的分类算法并根据特征信息对视频进行分类；
第一排序子模块，用于在各个类别下计算每个视频的综合分数，根据综合分数由高到低进行排序。
该第一排序子模块包括：
BaseScore(video)＝Hotness(video)×Freshness(video)；其中，BaseScore(video)代表视频的综合分数，Hotness(video)代表视频的热度，Freshness(video)代表视频的时新性。
在一个实施例中，该用户分析模块21进一步包括：第二获取子模块，确定子模块和归一化子模块。
第二获取子模块与确定子模块电连接，用于获取各个类别下的视频与用户标识对应的曝光量和点击量；
确定子模块与归一化子模块电连接，用于根据点击量与曝光量的比值确定用户标识对所述类别的喜好程度值，即：
Figure PCTCN2016088113-appb-000003
其中category表示类别，user表示用户，Click(user,category)表示用户user对category类别下的视频的点击量，Exposure(user,category)表示category类别下的视频在用户user登录时的曝光量；
归一化子模块，用于对所述类别的喜好程度值进行归一化处理，即：
Figure PCTCN2016088113-appb-000004
其中，NormaizeFavorite(user,category)代表归一化喜好程度值，MaxFavorite(user,category)代表用户user对各个类别category的最大喜好程度值。
在一个实施例中，该视频推送模块23进一步包括：第一拉取子模块，过滤子模块和第一推送子模块。
第一拉取子模块与过滤子模块电连接，用于根据登录的用户标识对各个类别的喜好程度值和排序的结果从各个分类下拉取视频；
过滤子模块与第一推送子模块电连接，用于在拉取的视频中过滤掉已对用户标识曝光过的视频；
第一推送子模块，用于将过滤后的视频推送至终端设备进行展示。
在一个实施例中，该视频推送模块23进一步包括：第二拉取子模块，第二排序子模块和第二推送子模块。
第二拉取子模块与第二排序子模块电连接，用于根据登录的用户标识对各个类别的喜好程度值和排序的结果从各个分类下拉取视频；
第二排序子模块与第二推送子模块电连接，用于对拉取的视频进行打分排序，并且排序结果中连续属于同一类别的视频数小于或等于预设数量；
第二推送子模块，用于将排序后的视频推送至终端设备进行展示。
此外，本申请实施例中可以通过硬件处理器(hardware processor)来实现上述各个功能模块。
本申请实施例还提供了一种视频推荐设备，所述设备包括：处理器和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：对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；根据浏览记录分析各个用户标识对各个类别的喜好程度值；根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；根据所述喜好程度值和所述排序的结果拉取各个分类下的视频并推送至终端设备进行展示。
所述视频的受欢迎程度为视频的综合分数，所述对视频进行分类，并根据视频的综合分数对各个类别下的视频进行排序包括：获取所述视频的特征信息；利用预设的分类算法并根据所述特征信息对视频进行分类；在各个类别下计算每个视频的综合分数，根据所述综合分数由高到低进行排序。
所述在各个类别下计算每个视频的综合分数包括：
BaseScore(video)＝Hotness(video)×Freshness(video)；其中，BaseScore(video)代表视频的综合分数，Hotness(video)代表视频的热度，Freshness(video)代表视频的时新性。
所述根据浏览记录分析各个用户标识对各个类别的喜好程度值包括：获取各个类别下的视频与所述用户标识对应的曝光量和点击量；根据所述点击量与曝光量的比值确定所述用户标识对所述类别的喜好程度值，即：
Figure PCTCN2016088113-appb-000005
其中category表示类别，user表示用户，Click(user,category)表示用户user对category类别下的视频的点击量，Exposure(user,category)表示category类别 下的视频在用户user登录时的曝光量；
对所述类别的喜好程度值进行归一化处理，即：
Figure PCTCN2016088113-appb-000006
其中，NormaizeFavorite(user,category)代表归一化喜好程度值，MaxFavorite(user,category)代表用户user对各个类别category的最大喜好程度值。
所述根据所述喜好程度值和所述排序的结果拉取各个分类下的视频并推送至终端设备进行展示包括：根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；在拉取的视频中过滤掉已对所述用户标识曝光过的视频；将过滤后的视频推送至终端设备进行展示。
所述根据所述喜好程度值和所述排序的结果拉取各个分类下的视频并推送至终端设备进行展示包括：根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；对拉取的视频进行打分排序，并且排序结果中连续属于同一类别的视频数小于或等于预设数量；将排序后的视频推送至终端设备进行展示。
以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。
此外，根据本申请的方法还可以被实现为由CPU执行的计算机程序，该计算机程序可以存储在计算机可读存储介质中。在该计算机程序被CPU执行时，执行本申请的方法中限定的上述功能。
此外，上述方法步骤以及系统单元也可以利用控制器以及用于存储使得控制器实现上述步骤或单元功能的计算机程序的计算机可读存储介质实现。
此外，应该明白的是，本文所述的计算机可读存储介质(例如，存储器)可以是易失性存储器或非易失性存储器，或者可以包括易失性存储器和非易失 性存储器两者。作为例子而非限制性的，非易失性存储器可以包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦写可编程ROM(EEPROM)或快闪存储器。易失性存储器可以包括随机存取存储器(RAM)，该RAM可以充当外部高速缓存存储器。作为例子而非限制性的，RAM可以以多种形式获得，比如同步RAM(DRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据速率SDRAM(DDR SDRAM)、增强SDRAM(ESDRAM)、同步链路DRAM(SLDRAM)以及直接RambusRAM(DRRAM)。所公开的方面的存储设备意在包括但不限于这些和其它合适类型的存储器。
本领域技术人员还将明白的是，结合这里的公开所描述的各种示例性逻辑块、模块、电路和算法步骤可以被实现为电子硬件、计算机软件或两者的组合。为了清楚地说明硬件和软件的这种可互换性，已经就各种示意性组件、方块、模块、电路和步骤的功能对其进行了一般性的描述。这种功能是被实现为软件还是被实现为硬件取决于具体应用以及施加给整个系统的设计约束。本领域技术人员可以针对每种具体应用以各种方式来实现所述的功能，但是这种实现决定不应被解释为导致脱离本申请的范围。
结合这里的公开所描述的各种示例性逻辑块、模块和电路可以利用被设计成用于执行这里所述功能的下列部件来实现或执行：通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或其它可编程逻辑器件、分立门或晶体管逻辑、分立的硬件组件或者这些部件的任何组合。通用处理器可以是微处理器，但是可替换地，处理器可以是任何传统处理器、控制器、微控制器或状态机。处理器也可以被实现为计算设备的组合，例如，DSP和微处理器的组合、多个微处理器、一个或多个微处理器结合DSP核、或任何其它这种配置。
结合这里的公开所描述的方法或算法的步骤可以直接包含在硬件中、由处理器执行的软件模块中或这两者的组合中。软件模块可以驻留在RAM存储器、快闪存储器、ROM存储器、EPROM存储器、EEPROM存储器、寄存器、硬盘、可移动盘、CD-ROM、或本领域已知的任何其它形式的存储介质中。示例性的存储介质被耦合到处理器，使得处理器能够从该存储介质中 读取信息或向该存储介质写入信息。在一个替换方案中，所述存储介质可以与处理器集成在一起。处理器和存储介质可以驻留在ASIC中。ASIC可以驻留在用户终端中。在一个替换方案中，处理器和存储介质可以作为分立组件驻留在服务器中。
在一个或多个示例性设计中，所述功能可以在硬件、软件、固件或其任意组合中实现。如果在软件中实现，则可以将所述功能作为一个或多个指令或代码存储在计算机可读介质上或通过计算机可读介质来传送。计算机可读介质包括计算机存储介质和通信介质，该通信介质包括有助于将计算机程序从一个位置传送到另一个位置的任何介质。存储介质可以是能够被通用或专用计算机访问的任何可用介质。作为例子而非限制性的，该计算机可读介质可以包括RAM、ROM、EEPROM、CD-ROM或其它光盘存储设备、磁盘存储设备或其它磁性存储设备，或者是可以用于携带或存储形式为指令或数据结构的所需程序代码并且能够被通用或专用计算机或者通用或专用处理器访问的任何其它介质。此外，任何连接都可以适当地称为计算机可读介质。例如，如果使用同轴线缆、光纤线缆、双绞线、数字用户线路(DSL)或诸如红外线、无线电和微波的无线技术来从网站、服务器或其它远程源发送软件，则上述同轴线缆、光纤线缆、双绞线、DSL或诸如红外先、无线电和微波的无线技术均包括在介质的定义。如这里所使用的，磁盘和光盘包括压缩盘(CD)、激光盘、光盘、数字多功能盘(DVD)、软盘、蓝光盘，其中磁盘通常磁性地再现数据，而光盘利用激光光学地再现数据。上述内容的组合也应当包括在计算机可读介质的范围内。
公开的仅为示例性实施例，但是应当注意，在不背离权利要求限定的本申请的范围的前提下，可以进行多种改变和修改。根据这里描述的公开实施例的方法权利要求的功能、步骤和/或动作不需以任何特定顺序执行。此外，尽管本申请的元素可以以个体形式描述或要求，但是也可以设想多个，除非明确限制为单数。
应当理解的是，在本文中使用的，除非上下文清楚地支持例外情况，单数形式“一个”(“a”、“an”、“the”)旨在也包括复数形式。还应当理解的是，在本文中使用的“和/或”是指包括一个或者一个以上相关联地列出的项目的 任意和所有可能组合。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
以上所述仅为本申请的较佳实施例，并不用以限制本申请，凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。
工业实用性
通过本申请提供的视频推荐方法、装置和设备，对视频进行分类排序并对用户喜好进行分析，根据登录的用户标识确定用户的喜好，推送相应类别的且排名靠前的视频至终端设备，在视频应用程序首页中展示推送的视频，从而使用户能够直接看到符合自己喜好的视频，由于推送的视频是各个类别中排序靠前的视频，因此推送视频的受欢迎程度也能够得到提升，而且免去了用户手动输入关键词进行搜索的操作，更加便于用户使用。

Claims (13)

  1. 一种视频推荐方法，其特征在于，所述方法包括：
    对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；
    根据浏览记录分析各个用户标识对各个类别的喜好程度值；
    根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；
    根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示。
  2. 根据权利要求1所述的方法，其特征在于，所述视频的受欢迎程度为所述视频的综合分数；
    所述对视频进行分类，并根据视频的综合分数对各个类别下的视频进行排序包括：
    获取所述视频的特征信息；
    利用预设的分类算法并根据所述特征信息对视频进行分类；
    在各个类别下计算每个视频的综合分数，根据所述综合分数由高到低进行排序。
  3. 根据权利要求2所述的方法，其特征在于，所述在各个类别下计算每个视频的综合分数包括：
    BaseScore(video)＝Hotness(video)×Freshness(video)；其中，BaseScore(video)代表视频的综合分数，Hotness(video)代表视频的热度，Freshness(video)代表视频的时新性。
  4. 根据权利要求1所述的方法，其特征在于，所述根据浏览记录分析各个用户标识对各个类别的喜好程度值包括：
    获取各个类别下的视频与所述用户标识对应的曝光量和点击量；
    根据所述点击量与曝光量的比值确定所述用户标识对所述类别的喜好 程度值，即
    Figure PCTCN2016088113-appb-100001
    其中category表示类别，user表示用户，Click(user,category)表示用户user对category类别下的视频的点击量，Exposure(user,category)表示category类别下的视频在用户user登录时的曝光量；
    对所述类别的喜好程度值进行归一化处理，即：
    Figure PCTCN2016088113-appb-100002
    其中，NormaizeFavorite(user,category)代表归一化喜好程度值，MaxFavorite(user,category)代表用户user对各个类别category的最大喜好程度值。
  5. 根据权利要求1所述的方法，其特征在于，所述根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示，包括：
    根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；
    在拉取的视频中过滤掉已对所述用户标识曝光过的视频；
    将过滤后的视频推送至终端设备进行展示。
  6. 根据权利要求1所述的方法，其特征在于，所述根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示，包括：
    根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；
    对拉取的视频进行打分排序，并且排序结果中连续属于同一类别的视频数小于或等于预设数量；
    将排序后的视频推送至终端设备进行展示。
  7. 一种视频推荐装置，其特征在于，所述装置包括：
    视频分类模块，用于对视频进行分类，并根据视频的受欢迎程度对各个 类别下的视频进行排序；
    用户分析模块，用于根据浏览记录分析各个用户标识对各个类别的喜好程度值；
    数据获取模块，用于根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好程度值，获取所述登录的用户标识对各个类别的喜好程度值；
    视频推送模块，用于根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示。
  8. 根据权利要求7所述的装置，其特征在于，所述视频的受欢迎程度为所述视频的综合分数；所述视频分类模块包括：
    第一获取子模块，用于获取所述视频的特征信息；
    分类子模块，用于利用预设的分类算法并根据所述特征信息对视频进行分类；
    第一排序子模块，用于在各个类别下计算每个视频的综合分数，根据所述综合分数由高到低进行排序。
  9. 根据权利要求8所述的装置，其特征在于，所述第一排序子模块包括：
    BaseScore(video)＝Hotness(video)×Freshness(video)；其中，BaseScore(video)代表视频的综合分数，Hotness(video)代表视频的热度，Freshness(video)代表视频的时新性。
  10. 根据权利要求7所述的装置，其特征在于，所述用户分析模块包括：
    第二获取子模块，用于获取各个类别下的视频与所述用户标识对应的曝光量和点击量；
    确定子模块，用于根据所述点击量与曝光量的比值确定所述用户标识对所述类别的喜好程度值，即
    Figure PCTCN2016088113-appb-100003
    其中category表示类别，user表示用户，Click(user,category)表示用户user对category类别下的视频的点击量，Exposure(user,category)表示category类别 下的视频在用户user登录时的曝光量；
    归一化子模块，用于对所述类别的喜好程度值进行归一化处理，即：
    Figure PCTCN2016088113-appb-100004
    其中，NormaizeFavorite(user,category)代表归一化喜好程度值，MaxFavorite(user,category)代表用户user对各个类别category的最大喜好程度值。
  11. 根据权利要求7所述的装置，其特征在于，所述视频推送模块包括：
    第一拉取子模块，用于根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；
    过滤子模块，用于在拉取的视频中过滤掉已对所述用户标识曝光过的视频；
    第一推送子模块，用于将过滤后的视频推送至所述终端设备进行展示。
  12. 根据权利要求7所述的装置，其特征在于，所述视频推送模块包括：
    第二拉取子模块，用于根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果从各个分类下拉取视频；
    第二排序子模块，用于对拉取的视频进行打分排序，并且排序结果中连续属于同一类别的视频数小于或等于预设数量；
    第二推送子模块，用于将排序后的视频推送至终端设备进行展示。
  13. 一种视频推荐设备，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：
    对视频进行分类，并根据视频的受欢迎程度对各个类别下的视频进行排序；
    根据浏览记录分析各个用户标识对各个类别的喜好程度值；
    根据登录终端设备的用户标识和所述各个用户标识对各个类别的喜好 程度值，获取所述登录的用户标识对各个类别的喜好程度值；
    根据所述登录的用户标识对各个类别的喜好程度值和所述排序的结果拉取各个分类下的视频并推送至所述终端设备进行展示。
PCT/CN2016/088113 2015-12-15 2016-07-01 视频推荐方法、装置和设备 WO2017101299A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/247,758 US20170169040A1 (en) 2015-12-15 2016-08-25 Method and electronic device for recommending video

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510938060.2A CN105893443A (zh) 2015-12-15 2015-12-15 视频推荐方法、装置和服务器
CN201510938060.2 2015-12-15

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/247,758 Continuation US20170169040A1 (en) 2015-12-15 2016-08-25 Method and electronic device for recommending video

Publications (1)

Publication Number Publication Date
WO2017101299A1 true WO2017101299A1 (zh) 2017-06-22

Family

ID=57002492

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/088113 WO2017101299A1 (zh) 2015-12-15 2016-07-01 视频推荐方法、装置和设备

Country Status (2)

Country Link
CN (1) CN105893443A (zh)
WO (1) WO2017101299A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109886719A (zh) * 2018-12-20 2019-06-14 平安科技（深圳）有限公司 基于网格的数据挖掘处理方法、装置和计算机设备
CN111026913A (zh) * 2019-12-10 2020-04-17 北京奇艺世纪科技有限公司 一种视频分发方法、装置、电子设备及存储介质
CN111143611A (zh) * 2019-12-31 2020-05-12 新疆联海创智信息科技有限公司 一种信息获取方法及装置
CN111241405A (zh) * 2020-01-15 2020-06-05 广东小天才科技有限公司 用户画像的标签获取方法、标签获取装置及终端设备
CN112333547A (zh) * 2020-11-20 2021-02-05 广州欢网科技有限责任公司 电视端短视频用户喜爱偏好推荐方法、装置和智能电视
CN112565797A (zh) * 2019-09-10 2021-03-26 北京达佳互联信息技术有限公司 一种页面请求处理方法、装置、电子设备及介质
CN112650884A (zh) * 2020-12-21 2021-04-13 四川长虹电器股份有限公司 基于深度学习的视频推送方法
CN112650931A (zh) * 2021-01-04 2021-04-13 杭州情咖网络技术有限公司 一种内容推荐方法
CN114245220A (zh) * 2020-09-09 2022-03-25 中国联合网络通信集团有限公司 在线视频评价方法、装置、计算机设备及存储介质
CN114595381A (zh) * 2022-02-22 2022-06-07 平安科技（深圳）有限公司 推荐方法、推荐装置、电子设备及存储介质
CN115334321A (zh) * 2022-08-10 2022-11-11 北京百度网讯科技有限公司 视频流的访问热度的获取方法、装置、电子设备及介质

Families Citing this family (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106372620B (zh) * 2016-09-29 2021-06-29 北京小米移动软件有限公司 视频信息的共享方法及装置
CN106792175A (zh) * 2016-12-22 2017-05-31 深圳Tcl数字技术有限公司 节目数据处理方法及系统
CN106993030A (zh) * 2017-03-22 2017-07-28 北京百度网讯科技有限公司 基于人工智能的信息推送方法和装置
CN106991192A (zh) * 2017-04-13 2017-07-28 武汉斗鱼网络科技有限公司 首页推荐提醒方法及装置
CN108733684A (zh) * 2017-04-17 2018-11-02 合信息技术（北京）有限公司 多媒体资源的推荐方法及装置
CN107071528A (zh) * 2017-04-20 2017-08-18 暴风集团股份有限公司 一种体育赛程的显示方法和显示装置
CN107249145B (zh) * 2017-05-05 2019-08-02 中广热点云科技有限公司 一种推送视频的方法
CN109151542B (zh) * 2017-06-28 2021-07-23 武汉斗鱼网络科技有限公司 处理违规直播间的方法、装置和设备及计算机可读存储介质
CN107748759A (zh) * 2017-09-21 2018-03-02 广州华多网络科技有限公司 视频推送方法及其设备
CN108259939B (zh) * 2017-12-25 2020-08-07 广州华多网络科技有限公司 新视频推送控制方法、装置及服务器
CN110019957A (zh) * 2017-12-27 2019-07-16 飞狐信息技术（天津）有限公司 一种视频推荐方法及装置
CN108600329B (zh) * 2018-03-29 2022-05-03 南京尚网网络科技有限公司 用于推送信息、展示信息的方法和设备
CN108595580B (zh) * 2018-04-17 2022-08-09 创新先进技术有限公司 新闻推荐方法、装置、服务器及存储介质
CN108712665B (zh) * 2018-05-18 2021-04-23 武汉斗鱼网络科技有限公司 一种直播列表的生成方法、装置、服务器及存储介质
CN109189951B (zh) * 2018-07-03 2021-09-03 南京尚网网络科技有限公司 一种多媒体资源推荐方法、设备及存储介质
CN109063001B (zh) * 2018-07-09 2021-06-04 北京小米移动软件有限公司 页面展示方法及装置
CN109120865A (zh) * 2018-09-26 2019-01-01 北京分钟时代科技有限公司 瀑布流视频生成方法
CN111078990A (zh) * 2018-10-18 2020-04-28 千寻位置网络有限公司 对资源链接进行编排推荐系统及方法
CN109658129A (zh) * 2018-11-22 2019-04-19 北京奇虎科技有限公司 一种用户画像的生成方法及装置
CN109635131B (zh) * 2018-12-17 2022-09-16 北京字节跳动网络技术有限公司 多媒体内容榜单显示方法、推送方法，装置及存储介质
CN109800324A (zh) * 2018-12-18 2019-05-24 北京达佳互联信息技术有限公司 视频数据推荐方法、装置、服务器及存储介质
CN111353088A (zh) * 2018-12-24 2020-06-30 中移（杭州）信息技术有限公司 一种电子资源推荐方法、装置和可读介质
CN110096613B (zh) * 2019-04-12 2021-07-20 北京奇艺世纪科技有限公司 一种视频推荐方法、装置、电子设备及存储介质
CN110611848A (zh) * 2019-09-30 2019-12-24 咪咕视讯科技有限公司 信息处理方法、系统、终端、服务器及可读存储介质
CN111246257B (zh) * 2020-03-17 2022-07-26 百度在线网络技术（北京）有限公司 视频推荐方法、装置、设备及存储介质
CN112565871A (zh) * 2020-11-06 2021-03-26 深圳市易平方网络科技有限公司 一种视频预加载方法、智能终端及存储介质
CN113179431A (zh) * 2021-04-19 2021-07-27 广州欢网科技有限责任公司 Tv视频应用智能无限瀑布流方法、装置和设备
CN113449147A (zh) * 2021-07-06 2021-09-28 乐视云计算有限公司 基于主题的视频推荐方法和装置
CN113657766A (zh) * 2021-08-18 2021-11-16 深圳华侨城创新研究院有限公司 一种基于游客多元数据的旅游景区欢乐指数的计量方法
CN113901258A (zh) * 2021-10-27 2022-01-07 北京达佳互联信息技术有限公司 一种视频推荐方法、装置、电子设备和存储介质
CN115878903B (zh) * 2023-02-21 2023-05-09 万链指数(青岛)信息科技有限公司 基于大数据的信息智能推荐方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102236669A (zh) * 2010-04-30 2011-11-09 腾讯科技（深圳）有限公司 推荐列表的生成方法和装置、媒体服务器、客户端及方法
EP2763421A1 (en) * 2013-02-01 2014-08-06 Krea Icerik Hizmetleri Ve Produksiyon Anonim Sirketi A personalized movie recommendation method and system
CN104216879A (zh) * 2013-05-29 2014-12-17 酷盛(天津）科技有限公司 视频质量挖掘系统及方法
CN104462573A (zh) * 2014-12-29 2015-03-25 北京奇艺世纪科技有限公司 一种视频搜索结果展示方法及装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102946566B (zh) * 2012-10-24 2015-07-01 北京奇虎科技有限公司 基于历史信息的视频推荐方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102236669A (zh) * 2010-04-30 2011-11-09 腾讯科技（深圳）有限公司 推荐列表的生成方法和装置、媒体服务器、客户端及方法
EP2763421A1 (en) * 2013-02-01 2014-08-06 Krea Icerik Hizmetleri Ve Produksiyon Anonim Sirketi A personalized movie recommendation method and system
CN104216879A (zh) * 2013-05-29 2014-12-17 酷盛(天津）科技有限公司 视频质量挖掘系统及方法
CN104462573A (zh) * 2014-12-29 2015-03-25 北京奇艺世纪科技有限公司 一种视频搜索结果展示方法及装置

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109886719A (zh) * 2018-12-20 2019-06-14 平安科技（深圳）有限公司 基于网格的数据挖掘处理方法、装置和计算机设备
CN112565797B (zh) * 2019-09-10 2022-08-05 北京达佳互联信息技术有限公司 一种页面请求处理方法、装置、电子设备及介质
CN112565797A (zh) * 2019-09-10 2021-03-26 北京达佳互联信息技术有限公司 一种页面请求处理方法、装置、电子设备及介质
CN111026913A (zh) * 2019-12-10 2020-04-17 北京奇艺世纪科技有限公司 一种视频分发方法、装置、电子设备及存储介质
CN111026913B (zh) * 2019-12-10 2024-04-23 北京奇艺世纪科技有限公司 一种视频分发方法、装置、电子设备及存储介质
CN111143611A (zh) * 2019-12-31 2020-05-12 新疆联海创智信息科技有限公司 一种信息获取方法及装置
CN111143611B (zh) * 2019-12-31 2024-01-16 新疆联海创智信息科技有限公司 一种信息获取方法及装置
CN111241405A (zh) * 2020-01-15 2020-06-05 广东小天才科技有限公司 用户画像的标签获取方法、标签获取装置及终端设备
CN114245220B (zh) * 2020-09-09 2023-05-16 中国联合网络通信集团有限公司 在线视频评价方法、装置、计算机设备及存储介质
CN114245220A (zh) * 2020-09-09 2022-03-25 中国联合网络通信集团有限公司 在线视频评价方法、装置、计算机设备及存储介质
CN112333547A (zh) * 2020-11-20 2021-02-05 广州欢网科技有限责任公司 电视端短视频用户喜爱偏好推荐方法、装置和智能电视
CN112650884A (zh) * 2020-12-21 2021-04-13 四川长虹电器股份有限公司 基于深度学习的视频推送方法
CN112650931A (zh) * 2021-01-04 2021-04-13 杭州情咖网络技术有限公司 一种内容推荐方法
CN112650931B (zh) * 2021-01-04 2023-05-30 杭州情咖网络技术有限公司 一种内容推荐方法
CN114595381A (zh) * 2022-02-22 2022-06-07 平安科技（深圳）有限公司 推荐方法、推荐装置、电子设备及存储介质
CN115334321A (zh) * 2022-08-10 2022-11-11 北京百度网讯科技有限公司 视频流的访问热度的获取方法、装置、电子设备及介质
CN115334321B (zh) * 2022-08-10 2024-04-09 北京百度网讯科技有限公司 视频流的访问热度的获取方法、装置、电子设备及介质

Also Published As

Publication number Publication date
CN105893443A (zh) 2016-08-24

Similar Documents

Publication Publication Date Title
WO2017101299A1 (zh) 视频推荐方法、装置和设备
US9253511B2 (en) Systems and methods for performing multi-modal video datastream segmentation
CN106331778B (zh) 视频推荐方法和装置
CN107124653B (zh) 电视用户画像的构建方法
US20160014482A1 (en) Systems and Methods for Generating Video Summary Sequences From One or More Video Segments
WO2017181612A1 (zh) 个性化视频推荐方法及装置
JP2020504475A (ja) ビデオデータ再生中の関連オブジェクトの提供
US9913001B2 (en) System and method for generating segmented content based on related data ranking
US20170169040A1 (en) Method and electronic device for recommending video
US20100100566A1 (en) Methods and Systems for Identifying the Fantasies of Users Based on Image Tagging
WO2014138305A1 (en) Systems and methods for providing user interactions with media
Choudhari et al. Video search engine optimization using keyword and feature analysis
US20130007057A1 (en) Automatic image discovery and recommendation for displayed television content
CN108197336B (zh) 一种视频查找的方法及装置
WO2013037223A1 (zh) 网络微博名人信息的推荐处理方法和装置
US11388561B2 (en) Providing a summary of media content to a communication device
US20090144226A1 (en) Information processing device and method, and program
WO2015182064A1 (en) Information processing device, information processing method, and program
US20150112814A1 (en) System and method for an integrated content publishing system
US20190014071A1 (en) Network information identification method and apparatus
US11036743B2 (en) Methods, systems, and media for presenting content organized by category
WO2016206035A1 (zh) 一种信息推荐方法及用户终端
TW201933231A (zh) 物品間關係探勘及推薦方法、裝置、計算設備、儲存媒體
TWI538491B (zh) 電視服務系統與提供影音服務的方法
CN110942070B (zh) 内容展示方法、装置、电子设备及计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16874363

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16874363

Country of ref document: EP

Kind code of ref document: A1