WO2014032492A1 - 媒体内容推荐方法及设备 - Google Patents
媒体内容推荐方法及设备 Download PDFInfo
- Publication number
- WO2014032492A1 WO2014032492A1 PCT/CN2013/080401 CN2013080401W WO2014032492A1 WO 2014032492 A1 WO2014032492 A1 WO 2014032492A1 CN 2013080401 W CN2013080401 W CN 2013080401W WO 2014032492 A1 WO2014032492 A1 WO 2014032492A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- media content
- candidate
- tag
- tags
- library
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
- G06F16/24578—Query processing with adaptation to user needs using ranking
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
- G06F16/437—Administration of user profiles, e.g. generation, initialisation, adaptation, distribution
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
Definitions
- the subject matter of the present disclosure relates to the field of personalized recommendation, and in particular, to a media content recommendation method and device.
- Background technique
- Song personalized recommendation is one of the most popular applications in the current personalized recommendation field. By finding the songs that the user likes and recommending to the user, the user can enrich the range of the songs and enhance the user's stickiness.
- the collaborative filtering recommendation method does not need to obtain the characteristics of the user or the song in advance, and only depends on the past behavior of the user (such as browsing the song, etc.), collects the feedback of the user on the song in the form of a score, and then calculates the similarity between the users. Degree, and then use the evaluation of other songs by the neighbors with higher user similarity to predict the preference of the target user to treat the song, and finally recommend the target user according to the degree of preference.
- this collaborative filtering recommendation method has at least a problem that the style of the recommended song is large and the recommended song is unpopular.
- the technical problem to be solved by the present invention is to provide a media content recommendation method and device, which can reduce the style difference of the recommended media content and avoid recommending unpopular media content.
- a media content recommendation method comprising the following steps:
- a media content recommendation device including: a calculation unit, configured to calculate a score of a media content tag in the media content library;
- a first selection unit configured to select a first threshold value media content label as a candidate media content label according to a sequence of highest to lowest scores
- a searching unit configured to search, for the candidate media content tag, the media content corresponding to the candidate media content tag from the media content library
- a second selection unit configured to select, according to the media content corresponding to the candidate media content label, a second threshold value of the media content according to the browsing quantity from highest to lowest as the recommended media content label to be recommended Media content;
- a recommendation unit configured to recommend, to the user, the media content to be recommended corresponding to the candidate media content label.
- the first threshold value media content tag may be selected as the candidate media content tag according to the highest to lowest score, and the media content is selected from the media content.
- the media content corresponding to the candidate media content label is found in the library, and the second threshold value media content is selected as the candidate media content label corresponding to the recommended media content, and the candidate is selected according to the highest to lowest page views.
- the media content to be recommended corresponding to the media content tag is recommended to the user. In this way, popular media content can be recommended for the user, and the style difference of the media content recommended to the user can be effectively reduced, and the recommended recommendation to the user is unpopular in the media.
- FIG. 1 is a flowchart of a media content recommendation method according to a first embodiment of the present invention
- FIG. 2 is a flowchart of a media content recommendation method according to a second embodiment of the present invention
- FIG. 3 is a flowchart of a media content recommendation method according to a third embodiment of the present invention
- FIG. 5 is a structural diagram of a media content recommendation device according to a fifth embodiment of the present invention.
- the embodiment of the invention provides a media content recommendation method and device, which can improve the accuracy of recommending media content for the user, and reduce the style difference of the media content recommended to the user, and avoid recommending to the user the unpopular media content, thereby greatly improving The user experience.
- the details are described below separately.
- FIG. 1 is a flow chart of a media content recommendation method according to a first embodiment of the present invention. As shown in FIG. 1, the media content recommendation method starts from step 101.
- a score for the media content tag in the media content library is calculated.
- the media content library can be used to store media content such as songs, videos, voices, pictures, and text.
- the media content in the media content library can be any one or combination of songs, videos, voices, pictures, and text.
- the media content tag may include at least eight: a theme, a mood, a rhythm, a melody, a scene, a timbre, an instrumental performance, and a style genre. That is to say, in the above step 101, the scores of the theme, mood, rhythm, melody, scene, timbre, instrumental performance, and style genre of each media content can be separately calculated.
- these media tags can be used without eight, and many of them are possible.
- the eight media content tags here are merely examples.
- step 101 can be implemented by the following sub-steps:
- the media content tag heat (the total number of media content in the media content library containing the media content tag / the total number of all media content in the media content library);
- the score of the media content tag ( ⁇ (user rating of the media content containing the media content tag) / media content tag popularity).
- the score of each of the media content tags in the media content library is calculated in the above manner, so that the reliability of the score of each media content tag is higher, so that the popularity of the subsequently recommended media content is higher, wherein The high popularity of media content indicates that media content is more popular.
- step 102 the first threshold value media content tag is selected as the candidate media content tag in descending order of the score.
- the foregoing first threshold value may be set according to actual needs, which is not specifically limited in the embodiment of the present invention.
- the present invention can rank the eight media content tags in a high-to-low order according to the theme, mood, rhythm melody, scene, timbre, instrumental performance, and style genre. Select the 5 highest media content tags as candidate media in-answer tags.
- step 103 for the candidate media content tags, the media content corresponding to the candidate media content tags is found from the media content library.
- the same media content tag can be owned by multiple media content, for example, the same voice can be owned by multiple songs at the same time. Therefore, after the first threshold value media content tag is selected as the candidate media content tag in step 102, the present invention can find a plurality of media content corresponding to each candidate media content tag from the media content library.
- step 104 for the media content corresponding to the candidate media content label, the second threshold value media content is selected as the media content to be recommended corresponding to the candidate media content label according to the highest to lowest page views.
- the foregoing second threshold value may be set according to actual needs, which is not specifically limited in the embodiment of the present invention.
- the embodiment of the present invention may separately select, for each of the 5 candidate media content tags, the media content tags.
- the media content is selected from the highest to the lowest, and the media content to be recommended corresponding to the candidate media content tag is selected.
- the amount of browsing of the media content can be understood as the number of listeners of the media content, and can also be understood as the number of clicks of the media content.
- step 105 can be
- 100 media content to be recommended is recommended to the user.
- the embodiment of the present invention may first sort each media content label in descending order of the score, and then perform step 102, which is not limited in the embodiment of the present invention. . By this sorting, step 102 can be made to select the first threshold value of the media content tag as a candidate media content tag.
- the embodiment of the present invention may filter the content and sort the media content corresponding to the remaining candidate media content labels according to the browsing amount from highest to lowest.
- the method of the present invention is not limited.
- the first filtering is recommended to the user again to reduce the user experience, and the step 104 can also select the second threshold value of the media content as the media content to be recommended corresponding to the candidate media content label.
- step 105 may be implemented by the following sub-steps:
- the media content to be recommended corresponding to each candidate media content label is sorted as a whole, and then ranked according to the ranking of each of the candidate media content labels from high to low, and then recommended to the user.
- the first threshold value media content tag may be selected as the candidate media according to the order of the score from highest to lowest. - - a content tag, and searching for the media content corresponding to the candidate media content tag from the media content library, and further selecting the second threshold value of the media content as the candidate media content tag according to the order of the pageview from highest to lowest.
- the media content to be recommended, and the media content to be recommended corresponding to the candidate media content tag are recommended to the user. In this way, it is possible to improve the accuracy of recommending popular media content for the user, and effectively reduce the style difference of the media content recommended to the user, and avoid recommending the media content that is unpopular to the user.
- FIG. 2 is a flow chart of a media content recommendation method according to a second embodiment of the present invention.
- the media content in the media content library is a song
- the media content recommendation method provided by the embodiment of the present invention is executed by the media content server.
- the media content recommendation method may include the following steps.
- the media content server records the user score for each song in the media content library.
- the media content server finds a song tag owned by the user's scored song in the media content library.
- the media content server calculates a score for each of the song tags in the media content library.
- step 203 can be specifically implemented by the following sub-steps:
- the song tag heat (the total number of songs in the media content library containing song tags / the total number of all songs in the media content library);
- the score of the song tag ( ⁇ (the user of the song containing the song tag scores) / the song heat).
- the score of each song tag in the media content library is calculated by the above manner, so that the score of each song tag is highly reliable, so that the heat of the subsequent recommended song is higher.
- the media content server ranks each song in descending order of highest score. - - Sort and select 5 (ie the first threshold) song labels as candidate song labels in descending order of score.
- the media content server may select melody, scene, timbre, instrumental performance, and style genre 5 (i.e., first threshold) song labels as candidate song labels in descending order of score.
- step 205 the media content server searches for a song corresponding to each candidate song tag from the media content library for each candidate song tag.
- each candidate song tag can have up to several hundred or even hundreds of songs.
- the media content server filters out the songs that have been browsed by the user or have been pushed to the user the previous day for the songs corresponding to each candidate song tag, and the remaining songs are as high as the number of listeners. Sorting in a low order, selecting 20 (ie, the second threshold) songs in descending order of the number of listeners as the songs to be recommended corresponding to the song labels of the candidate.
- the media content server for a song that has been browsed by a user, not only records the correspondence between the identifier (ID) of the song and the user ID (such as a QQ account), and the recorded song is recommended to the user. Time, therefore, the media content server can easily determine the song that has been viewed by the user or has been pushed to the user the previous day, and the media content server can filter out the user who has browsed or has been pushed to the user the previous day. song.
- ID the identifier
- the media content server sorts the songs to be recommended corresponding to each candidate song tag in descending order of the browsing amount, and sorts the songs to be recommended corresponding to each candidate song tag as a whole. Then, the scores of each candidate song tag are sorted in descending order, thereby obtaining 100 songs and recommending to the user.
- each song tag in the media content library after calculating the score of each song tag in the media content library, five song tags may be selected as the candidate song tags in descending order of the score, and the media content library is searched.
- the song corresponding to each candidate song tag is further selected according to the number of listeners from high to low, 20 songs are selected as the songs to be recommended corresponding to the candidate song tags, and the songs to be recommended corresponding to each candidate song tag are selected.
- Recommended for users In this way, popular media content can be recommended for the user, and the style difference of the media content recommended to the user can be effectively reduced, and the media content recommended to the user can be avoided.
- FIG. 3 is a flowchart of a media content recommendation method according to a third embodiment of the present invention.
- the media content in the media content library is a video
- the media content server performs the media content recommendation method provided by the embodiment of the present invention.
- the media content recommendation method may include the following steps.
- the media content server records a user rating for each video in the media content library.
- the media content server finds a video tag owned by the video scored by the user in the media content library.
- the media content server calculates a score for each of the video tags in the media content library.
- step 303 can be specifically implemented by the following sub-steps:
- the video tag heat (the total number of videos in the media content library containing video tags / the total number of videos in the media library);
- the score of the video tag ( ⁇ (the user of the video containing the video tag scores) / the video heat).
- the calculation of the score of each video tag in the media content library by the above manner can make the score of each video tag have higher credibility, so that the heat of the subsequent recommended video is higher.
- the media content server sorts each video tag in descending order of scores, and selects 3 (ie, the first threshold) video tags as candidate video tags in descending order of score. .
- the media content server may select the theme, mood, and style genre 3 (ie, the first threshold) video tags as candidate video tags in descending order of score.
- the theme, mood, and style genre 3 ie, the first threshold
- the media content server searches for a video corresponding to each candidate video tag from the media content library for each candidate video tag.
- each candidate video tag can have up to dozens or even hundreds of videos.
- the media content server first videos for each candidate video tag. - - Filters out videos that have been viewed by the user or have been pushed to the user the previous day, and sorts the remaining videos in descending order of clicks, in descending order of clicks 20 (ie, the second threshold) video is used as the video to be recommended corresponding to the candidate video tag.
- the media content server not only records the correspondence between the identifier (ID) of the video and the user ID (such as the user registration account), and the video recommendation is recorded to the video.
- ID the identifier
- the media content server can easily determine the video that has been viewed by the user or has been pushed to the user the previous day, and the media content server can filter out the information that has been viewed by the user or has been pushed the day before. User's video.
- the media content server sorts the to-be-recommended videos corresponding to each candidate video tag according to the browsing amount from high to low, and selects the to-be-recommended video corresponding to each candidate video tag as a whole. Then, the scores of each candidate video tag are sorted in descending order, thereby obtaining 60 videos and recommending to the user.
- three video tags may be selected as the candidate video tags in descending order of the score, and the media content library is searched.
- the video corresponding to each candidate video tag is further selected, and the video with the highest click volume is selected as the video to be recommended corresponding to the candidate video tag, and each candidate video tag is correspondingly selected according to the order of the number of clicks.
- the recommended videos are recommended to the user. In this way, the style difference of the video recommended to the user can be effectively reduced, and the video recommended to the user to be unpopular can be avoided, thereby improving the user experience.
- FIG. 4 is a structural diagram of a media content recommendation device according to a fourth embodiment of the present invention.
- the media content recommendation device shown in FIG. 4 may be a media content server or a smart platform with the media content recommendation capability, which is not limited in the embodiment of the present invention.
- the media content recommendation device may include:
- the calculating unit 401 calculates a score of the media content tag in the media content library.
- the media content in the media content library may be any one or a combination of songs, videos, voices, pictures, and texts.
- the media content tag can include at least eight: theme, mood, rhythm, melody, scene, timbre, instrumental performance and style genre. That is to say, the calculation unit 401 can separately calculate the scores of the theme, mood, rhythm, melody, scene, timbre, instrumental performance, and style genre of each media content in the media content library. - The first selection unit 402 selects the first threshold value media content tag as the candidate media content tag in descending order of the score.
- the searching unit 403 searches for the media content corresponding to the candidate media content tag from the media content library for the candidate media content tag.
- the second selecting unit 404 selects, according to the media content of the candidate media content tags, the second threshold value media content as the candidate media content tag corresponding to the media content to be recommended according to the browsing amount from high to low. Household.
- FIG. 5 is a structural diagram of a media content recommendation device according to a fifth embodiment of the present invention.
- the media content recommendation device shown in FIG. 5 is optimized by the media content recommendation device shown in FIG. 4.
- the computing unit 401 includes the following submodules:
- the first module 4011 calculates, for the media content tag in the media content library, a ratio of the total number of media content in the media content library that contains the media content tag to the total number of media content in the media content library, to obtain the media content tag.
- the second module 4012 calculates a ratio of the user score of the media content containing the media content tag in the media content library to the media content tag popularity, to obtain a score of the media content tag, and provides the score to the first selection unit 402.
- the recommendation unit 405 includes the following submodules:
- the third module 4051 is configured to sort the media content to be recommended corresponding to each candidate media content label in descending order of the browsing amount;
- the fourth module 4052 sorts the media content to be recommended corresponding to each candidate media content label, and then sorts the scores of each candidate media content label from high to low, and then recommends to the user.
- the first selection list 402 further selects the first threshold value media content label according to the order of the score from highest to lowest. Before the candidate media content tag, each media content tag is scored from high to - - Sort in low order.
- the second selection unit 404 is still in the media content corresponding to each candidate media content tag, according to the pageview amount from high to Before the second threshold value media content is selected as the media content to be recommended corresponding to the candidate media content label, the media content that has been browsed or pushed in the media content corresponding to the candidate media content label is filtered out. And sorting the media content corresponding to the remaining candidate media content tags in order of high to low page views.
- the media content recommendation device shown in FIG. 4 and FIG. 5 can recommend media content that is popular and conforms to the user's taste for the user, and effectively reduces the style difference of the media content recommended to the user, and avoids recommending the media content that is unpopular to the user. Thereby greatly improving the user experience.
- the media content recommendation method shown in FIG. 1 can be performed by each unit in the media content recommendation device shown in FIG.
- steps 101 to 105 shown in Fig. 1 can be performed by the calculation unit 401, the first selection unit 402, the search unit 403, the second selection unit 404, and the recommendation unit 405 shown in Fig. 4.
- each unit in the media content recommendation device shown in FIG. 4 may be separately or entirely combined into one or several additional units, or some of the units may be further Split into multiple units that are functionally smaller.
- the computing unit 401 can be implemented by the first module 4011 and the second module 4012
- the recommending unit 405 can be implemented by the third module 4051, the fourth module 4052, etc., which can achieve the same operation without affecting Implementation of the technical effects of the embodiments of the present invention.
- the hardware of the device is completed, and the program can be stored in a computer readable storage medium and executed by at least one processor.
- the storage medium may include: a flash drive, a read-only memory (ROM), and a random access device (random access)
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)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Signal Processing (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Description
一 一
媒体内容推荐方法及设备 本专利申请要求 2012 年 8 月 28 曰提交的中国专利申请号为 201210309267.X, 发明名称为 "一种媒体内容推荐方法及设备" 的优先权, 该 申请的全文以引用的方式并入本申请中。 技术领域
本发明公开的主题内容涉及个性化推荐领域,具体涉及一种媒体内容推荐 方法及设备。 背景技术
歌曲个性化推荐是当前个性化推荐领域最热门的应用之一。通过寻找用户 喜欢的歌曲并推荐给用户, 可以丰富用户的听歌范围, 增强用户粘性。
目前, 业界一般采用协同过滤推荐方法来为用户推荐歌曲。 其中, 协同过 滤推荐方法不需要事先获得用户或歌曲的特征, 只依赖于用户过去的行为(如 对歌曲的浏览等), 以评分的形式收集用户对歌曲的反馈, 接着计算用户之间 的相似度, 然后利用用户相似度较高的邻居对其他歌曲的评价, 来预测目标用 户对待定歌曲的喜好程度, 最后根据这一喜好程度来对目标用户进行推荐。
但是, 这种协同过滤推荐方法至少存在推荐歌曲的风格差异大,推荐歌曲 冷门的问题。 发明内容
本发明所要解决的技术问题是提供一种媒体内容推荐方法及设备,能够降 低所推荐的媒体内容的风格差异, 避免推荐冷门的媒体内容。
有鉴于此, 根据本发明的一个方面, 提供一种媒体内容推荐方法, 该媒体 内容推荐方法包括以下步骤:
计算媒体内容库中的媒体内容标签的得分;
按照得分从高到低的顺序选取第一门限值个媒体内容标签作为候选的媒 体内容标签;
针对所述候选的媒体内容标签,从所述媒体内容库中查找出所述候选的媒 体内容标签对应的媒体内容;
- - 针对所述候选的媒体内容标签对应的媒体内容,按照浏览量从高到低的顺 序选取第二门限值个媒体内容作为所述候选的媒体内容标签对应的待推荐媒 体内容; 根据本发明的另一方面, 还提供一种媒体内容推荐设备, 包括: 计算单元, 用于计算媒体内容库中的媒体内容标签的得分;
第一选择单元,用于按照得分从高到低的顺序选取第一门限值个媒体内容 标签作为候选的媒体内容标签;
查找单元, 用于针对所述候选的媒体内容标签,从所述媒体内容库中查找 出所述候选的媒体内容标签对应的媒体内容;
第二选择单元, 用于针对所述候选的媒体内容标签对应的媒体内容,按照 浏览量从高到低的顺序选取第二门限值个媒体内容作为所述候选的媒体内容 标签对应的待推荐媒体内容;
推荐单元 ,用于将所述候选的媒体内容标签对应的待推荐媒体内容推荐给 用户。
本发明实施例中,在计算媒体内容库中的媒体内容标签的得分之后, 可以 按照得分从高到低的顺序选取第一门限值个媒体内容标签作为候选的媒体内 容标签, 并且从媒体内容库中查找出候选的媒体内容标签对应的媒体内容, 进 一步按照浏览量从高到低的顺序选取第二门限值个媒体内容作为候选的媒体 内容标签对应的待推荐媒体内容,以及将候选的媒体内容标签对应的待推荐媒 体内容推荐给用户。 通过这种方式, 可以为用户推荐热门的媒体内容, 并且有 效地降低推荐给用户的媒体内容的风格差异, 避免推荐给用户冷门的媒体内
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施 例中所需要使用的附图作简单地介绍, 显而易见地, 下面描述中的附图仅仅是 本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性 的前提下, 还可以根据这些附图获得其他的附图。
图 1是本发明第一实施例提供的一种媒体内容推荐方法的流程图;
- - 图 2是本发明第二实施例提供的一种媒体内容推荐方法的流程图; 图 3是本发明第三实施例提供的一种媒体内容推荐方法的流程图; 图 4是本发明第四实施例提供的一种媒体内容推荐设备的结构图; 图 5是本发明第五实施例提供的一种媒体内容推荐设备的结构图。 具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清 楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部分实施例, 而不是 全部的实施例。基于本发明中的实施例, 本领域普通技术人员在没有做出创造 性劳动前提下所获得的所有其他实施例, 都属于本发明保护的范围。
本发明实施例提供了一种媒体内容推荐方法及设备,可以提高为用户推荐 媒体内容的准确性, 并且降低推荐给用户的媒体内容的风格差异,避免推荐给 用户冷门的媒体内容, 从而大大提高了用户体验。 以下分别进行详细说明。
请参阅图 1 , 图 1是本发明第一实施例提供的一种媒体内容推荐方法的流 程图。 如图 1所示, 该媒体内容推荐方法从步骤 101开始。
在步骤 101, 计算媒体内容库中的媒体内容标签的得分。
本发明中, 媒体内容库可以用于存储歌曲、 视频、 语音、 图片以及文本等 媒体内容。 换言之, 媒体内容库中的媒体内容可以是歌曲、 视频、 语音、 图片 以及文本中的任意一种或几种的组合。
以音乐文件为例, 在实际应用中, 媒体内容标签至少可以包括 8个: 即主 题、 心情、 节奏、 旋律、 场景、 音色、 乐器演奏以及风格流派。 也即是说, 上 述步骤 101中可以分别计算各媒体内容的主题、 心情、 节奏、 旋律、 场景、 音 色、 乐器演奏以及风格流派的得分。
需要说明的是, 实际应用中这些媒体标签可以不用 8个, 多少个都是可以 的, 此处的 8个媒体内容标签仅仅是举例说明而已。
作为一种可选的实施方式, 上述步骤 101 具体可以通过以下子步骤来实 现:
1 ) 、 针对媒体内容库中的媒体内容标签, 计算媒体内容库中含有该媒体 内容标签的媒体内容总数量与媒体内容库中所有媒体内容总数量的比值,以获 得媒体内容标签热度;
- - 即, 媒体内容标签热度 = (媒体内容库中含有该媒体内容标签的媒体内容 总数量 /媒体内容库中所有媒体内容总数量 ) ;
2 ) 、 计算媒体内容库中含有该媒体内容标签的媒体内容的用户打分总和 与该媒体内容标签热度的比值, 以获得该媒体内容标签的得分。
即, 媒体内容标签的得分 = (∑ (含有该媒体内容标签的媒体内容的用户 打分) /媒体内容标签热度) 。
其中, 通过上述方式来计算媒体内容库中的每一个媒体内容标签的得分, 可以使得每一个媒体内容标签的得分的可信度较高 ,从而使得后续推荐的媒体 内容的热度较高, 其中, 媒体内容的热度较高说明媒体内容较受欢迎。
在步骤 102, 按照得分从高到低的顺序选取第一门限值个媒体内容标签作 为候选的媒体内容标签。
本发明实施例中, 上述的第一门限值可以根据实际需要进行设置, 本发明 实施例不作具体限定。 举例来说, 当第一门限值设置为 5时, 本发明可以根据 主题、 心情、 节奏旋律、 场景、 音色、 乐器演奏以及风格流派等 8个媒体内容 标签的得分从高到低的顺序,选择 5个最高的媒体内容标签作为候选的媒体内 答标签。
在步骤 103 , 针对候选的媒体内容标签, 从媒体内容库中查找出候选的媒 体内容标签对应的媒体内容。
在实际应用中, 同一个媒体内容标签可以被多个媒体内容所拥有, 例如同 一个音色可以同时被多首歌曲所拥有。 因此,在步骤 102选取出第一门限值个 媒体内容标签作为候选的媒体内容标签之后,本发明可以从媒体内容库中查找 出每一个候选的媒体内容标签对应的多个媒体内容。
在步骤 104, 针对候选的媒体内容标签对应的媒体内容, 按照浏览量从高 到低的顺序选取第二门限值个媒体内容作为该候选的媒体内容标签对应的待 推荐媒体内容。
本发明实施例中, 上述的第二门限值可以根据实际需要进行设置, 本发明 实施例不作具体限定。 举例来说, 当第一门限值设置为 5 , 第二门限值设置为 20时, 本发明实施例可以针对这 5个候选的媒体内容标签中的每一个候选的 媒体内容标签, 分别按照浏览量从高到低的顺序选取 20个媒体内容作为该候 选的媒体内容标签对应的待推荐媒体内容。
本发明实施例中, 媒体内容的浏览量可以理解为是媒体内容的听众数量, 也可以理解为是媒体内容的点击数量。 举例来说, 当第一门限值设置为 5 , 第二门限值设置为 20时, 通过上述 步骤 104每一个候选的媒体内容标签对应的媒体内容中有 20个媒体内容被选 取作为该候选的媒体内容标签对应的待推荐媒体内容。此时通过步骤 105可以
100个待推荐媒体内容推荐给用户。
作为一种可选的实施方式, 本发明实施例在上述步骤 102之前, 可以先将 每一个媒体内容标签按照得分从高到低的顺序进行排序, 然后再执行步骤 102, 本发明实施例不作限定。 通过这种排序方式, 可以使得步骤 102选取第 一门限值个媒体内容标签作为候选的媒体内容标签更加便捷。
作为一种可选的实施方式, 本发明实施例在上述步骤 104之前, 可以先滤 容,并将剩下的候选的媒体内容标签对应的媒体内容按照浏览量从高到低的顺 序进行排序之后, 再执行步骤 104, 本发明实施例不作限定。 通过这种先滤除 再次推荐给用户, 降低用户体验, 而且还可以使步骤 104选取第二门限值个媒 体内容作为该候选的媒体内容标签对应的待推荐媒体内容更加便捷。
作为一种可选的实施方式, 上述步骤 105 具体可以通过以下子步骤来实 现:
1 ) 、 将每一个候选的媒体内容标签对应的待推荐媒体内容按照浏览量从 高到低的顺序进行排序;
2 ) 、 将排序好的每一个候选的媒体内容标签对应的待推荐媒体内容作为 一个整体,再按照每一个所述候选的媒体内容标签的得分从高到低的顺序进行 排序后推荐给用户。
其中,通过这种二次排序的方式, 可以确保首先推荐给用户的是最热门的 媒体内容标签对应的最热门的媒体内热, 从而可以进一步提高用户体验。
在图 1所示的方法,在计算媒体内容库中的媒体内容标签的得分之后, 可 以按照得分从高到低的顺序选取第一门限值个媒体内容标签作为候选的媒体
- - 内容标签, 并且从媒体内容库中查找出候选的媒体内容标签对应的媒体内容, 进一步按照浏览量从高到低的顺序选取第二门限值个媒体内容作为候选的媒 体内容标签对应的待推荐媒体内容,以及将候选的媒体内容标签对应的待推荐 媒体内容推荐给用户。通过这种方式, 可以为提高为用户推荐热门媒体内容的 准确性, 并且有效地降低推荐给用户的媒体内容的风格差异,避免推荐给用户 冷门的媒体内容。
请参阅图 2, 图 2是本发明第二实施例提供的一种媒体内容推荐方法的流 程图。 在图 2所示的方法中, 假设媒体内容库中的媒体内容都是歌曲, 并且由 媒体内容服务器来执行本发明实施例提供的媒体内容推荐方法。 如图 2所示, 该媒体内容推荐方法可以包括以下步骤。
在步骤 201 , 媒体内容服务器记录媒体内容库中的每一首歌曲的用户打 分。
在步骤 202, 媒体内容服务器找到媒体内容库中被用户打分的歌曲所拥有 的歌曲标签。
在步骤 203 , 媒体内容服务器计算媒体内容库中的每一个歌曲标签的得 分。
其中, 上述步骤 203具体可以通过以下子步骤来实现:
1 ) 、 针对媒体内容库中的每一个歌曲标签, 计算媒体内容库中含有该歌 曲标签的歌曲总数量与媒体内容库中所有歌曲总数量的比值,以获得该歌曲标 签热度;
即, 该歌曲标签热度= (媒体内容库中含有歌曲标签的歌曲总数量 /媒体内 容库中所有歌曲总数量) ;
2 ) 、 计算媒体内容库中含有该歌曲标签的媒体内容的用户打分总和与该 歌曲标签热度的比值, 以获得该歌曲标签的得分。
即, 该歌曲标签的得分 = (∑ (含有该歌曲标签的歌曲的用户打分) /该歌 曲热度) 。
其中,通过上述方式来计算媒体内容库中的每一个歌曲标签的得分, 可以 使得每一个歌曲标签的得分的可信度较高,从而使得后续推荐的歌曲的热度较 高。
在步骤 204, 媒体内容服务器将每一个歌曲标签按照得分从高到低的顺序
- - 进行排序, 并按照得分从高到低的顺序选取 5 (即第一门限值)个歌曲标签作 为候选的歌曲标签。
举例来说,媒体内容服务器可以按照得分从高到低的顺序选取旋律、场景、 音色、 乐器演奏以及风格流派这 5 (即第一门限值)个歌曲标签作为候选的歌 曲标签。
在步骤 205 , 媒体内容服务器针对每一个候选的歌曲标签, 从媒体内容库 中查找出每一个候选的歌曲标签对应的歌曲。
其中,每一个候选的歌曲标签对应的歌曲可以多达几十百首、甚至数百首。 在步骤 206, 媒体内容服务器针对每一个候选的歌曲标签对应的歌曲, 先 滤除掉已被用户浏览过或前一天已推送过给用户的歌曲,并将剩下的歌曲按照 听众数从高到低的顺序进行排序, 按照听众数从高到低的顺序选取 20 (即第 二门限值 )个歌曲作为该候选的歌曲标签对应的待推荐歌曲。
本发明实施例中,对于已被用户浏览过的歌曲,媒体内容服务器不仅记录 了该歌曲的标识(ID )与用户 ID (如 QQ账号)之间的对应关系, 该记录了 该歌曲推荐给用户的时间, 因此,媒体内容服务器可以轻易地确定出已被用户 浏览过或前一天已推送过给用户的歌曲,进而媒体内容服务器可以滤除掉已被 用户浏览过或前一天已推送过给用户的歌曲。
在步骤 207, 媒体内容服务器将每一个候选的歌曲标签对应的待推荐歌曲 按照浏览量从高到低的顺序进行排序,并将排序好的每一个候选的歌曲标签对 应的待推荐歌曲作为一个整体,再按照每一个候选的歌曲标签的得分从高到低 的顺序进行排序, 从而得到 100首歌曲, 并推荐给用户。
在图 2所示的方法, 在计算媒体内容库中的每一个歌曲标签的得分之后, 可以按照得分从高到低的顺序选取 5个歌曲标签作为候选的歌曲标签,并且从 媒体内容库中查找出每一个候选的歌曲标签对应的歌曲,进一步按照听众数从 高到低的顺序选取 20个歌曲作为该候选的歌曲标签对应的待推荐歌曲, 以及 将每一个候选的歌曲标签对应的待推荐歌曲推荐给用户。通过这种方式, 可以 为用户推荐热门的媒体内容,并且有效地降低推荐给用户的媒体内容的风格差 异, 避免推荐给用户冷门的媒体内容。
请参阅图 3 , 图 3是本发明第三实施例提供的一种媒体内容推荐方法的流 程图。 在图 3所示的方法中, 假设媒体内容库中的媒体内容都是视频, 并且由
- - 媒体内容服务器来执行本发明实施例提供的媒体内容推荐方法。 如图 2所示, 该媒体内容推荐方法可以包括以下步骤。
在步骤 301 , 媒体内容服务器记录媒体内容库中的每一个视频的用户打 分。
在步骤 302 , 媒体内容服务器找到媒体内容库中被用户打分的视频所拥有 的视频标签。
在步骤 303 , 媒体内容服务器计算媒体内容库中的每一个视频标签的得 分。
其中, 上述步骤 303具体可以通过以下子步骤来实现:
1 ) 、 针对媒体内容库中的每一个视频标签, 计算媒体内容库中含有该视 频标签的视频总数量与媒体内容库中所有视频总数量的比值,以获得该视频标 签热度;
即, 该视频标签热度= (媒体内容库中含有视频标签的视频总数量 /媒体内 容库中所有视频总数量) ;
2 ) 、 计算媒体内容库中含有该视频标签的媒体内容的用户打分总和与该 视频标签热度的比值, 以获得该视频标签的得分。
即, 该视频标签的得分 = (∑ (含有该视频标签的视频的用户打分) /该视 频热度) 。
其中,通过上述方式来计算媒体内容库中的每一个视频标签的得分, 可以 使得每一个视频标签的得分的可信度较高 ,从而使得后续推荐的视频的热度较 高。
在步骤 304 , 媒体内容服务器将每一个视频标签按照得分从高到低的顺序 进行排序, 并按照得分从高到低的顺序选取 3 (即第一门限值)个视频标签作 为候选的视频标签。
举例来说,媒体内容服务器可以按照得分从高到低的顺序选取主题、 心情 以及风格流派这 3 (即第一门限值)个视频标签作为候选的视频标签。
在步骤 305 , 媒体内容服务器针对每一个候选的视频标签, 从媒体内容库 中查找出每一个候选的视频标签对应的视频。
其中, 每一个候选的视频标签对应的视频可以多达几十个、 甚至数百个。 在步骤 306 , 媒体内容服务器针对每一个候选的视频标签对应的视频, 先
- - 过滤除掉已被用户浏览过或前一天已推送过给用户的视频,并将剩下的视频按 照点击数从高到低的顺序进行排序,并按照点击数从高到低的顺序选取 20(即 第二门限值 ) 个视频作为该候选的视频标签对应的待推荐视频。
本发明实施例中,对于已被用户浏览过的视频,媒体内容服务器不仅记录 了该视频的标识(ID )与用户 ID (如用户注册账号)之间的对应关系, 该记 录了该视频推荐给用户的时间, 因此,媒体内容服务器可以轻易地确定出已被 用户浏览过或前一天已推送过给用户的视频,进而媒体内容服务器可以滤除掉 已被用户浏览过或前一天已推送过给用户的视频。
在步骤 307, 媒体内容服务器将每一个候选的视频标签对应的待推荐视频 按照浏览量从高到低的顺序进行排序,并将排序好的每一个候选的视频标签对 应的待推荐视频作为一个整体,再按照每一个候选的视频标签的得分从高到低 的顺序进行排序, 从而得到 60个视频, 并推荐给用户。
在图 3所示的方法, 在计算媒体内容库中的每一个视频标签的得分之后, 可以按照得分从高到低的顺序选取 3个视频标签作为候选的视频标签,并且从 媒体内容库中查找出每一个候选的视频标签对应的视频,进一步按照点击数从 高到低的顺序选取 20个点击量最高的视频作为该候选的视频标签对应的待推 荐视频, 以及将每一个候选的视频标签对应的待推荐视频推荐给用户。通过这 种方式, 可以有效地降低推荐给用户的视频的风格差异,避免推荐给用户冷门 的视频, 从而可以提高了用户体验。
请参阅图 4, 图 4是本发明第四实施例提供的一种媒体内容推荐设备的结 构图。其中, 图 4所示的媒体内容推荐设备在物理实现上可以是媒体内容服务 器, 或者是其他具备媒体内容推荐能力的智能平台, 本发明实施例不作限定。 如图 4所示, 该媒体内容推荐设备可以包括:
计算单元 401 , 计算媒体内容库中的媒体内容标签的得分。
其中, 媒体内容库中的媒体内容可以是歌曲、 视频、 语音、 图片以及文本 中的任意一种或几种的组合。
其中, 媒体内容标签至少可以包括 8个: 即主题、 心情、 节奏、 旋律、 场 景、 音色、 乐器演奏以及风格流派。 也即是说, 计算单元 401可以分别计算媒 体内容库中各媒体内容的主题、 心情、 节奏、 旋律、 场景、 音色、 乐器演奏以 及风格流派的得分。
- - 第一选择单元 402, 按照得分从高到低的顺序选取第一门限值个媒体内容 标签作为候选的媒体内容标签。
查找单元 403 , 针对候选的媒体内容标签, 从媒体内容库中查找出候选的 媒体内容标签对应的媒体内容。
第二选择单元 404, 针对候选的媒体内容标签对应的媒体内容, 按照浏览 量从高到低的顺序选取第二门限值个媒体内容作为候选的媒体内容标签对应 的待推荐媒体内容。 户。
请一并参阅图 5 , 图 5本发明第五实施例提供的一种媒体内容推荐设备的 结构图。其中, 图 5所示的媒体内容推荐设备是由图 4所示的媒体内容推荐设 备进行优化得到的。
作为一种可选的实施方式,在图 5所示的媒体内容推荐设备中,计算单元 401包括以下子模块:
第一模块 4011 , 针对媒体内容库中的媒体内容标签, 计算该媒体内容库 中含有该媒体内容标签的媒体内容总数量与媒体内容库中所有媒体内容总数 量的比值, 以获得该媒体内容标签热度;
第二模块 4012, 计算媒体内容库中含有该媒体内容标签的媒体内容的用 户打分总和与该媒体内容标签热度的比值, 以获得该媒体内容标签的得分, 并 提供给第一选择单元 402。
作为一种可选的实施方式,在图 5所示的媒体内容推荐设备中,推荐单元 405包括以下子模块:
第三模块 4051 , 将每一个候选的媒体内容标签对应的待推荐媒体内容按 照浏览量从高到低的顺序进行排序;
第四模块 4052, 将排序好的每一个候选的媒体内容标签对应的待推荐媒 体内容作为一个整体,再按照每一个候选的媒体内容标签的得分从高到低的顺 序进行排序后推荐给用户。
作为一种可选的实施方式, 在图 4、 图 5所示的媒体内容推荐设备中, 第一选择单 402还在按照得分从高到低的顺序选取第一门限值个媒体内 容标签作为候选的媒体内容标签之前,将每一个媒体内容标签按照得分从高到
- - 低的顺序进行排序。
作为一种可选的实施方式, 在图 4、 图 5所示的媒体内容推荐设备中, 第二选择单元 404还在针对每一个候选的媒体内容标签对应的媒体内容, 按照浏览量从高到低的顺序选取第二门限值个媒体内容作为候选的媒体内容 标签对应的待推荐媒体内容之前,滤除候选的媒体内容标签对应的媒体内容中 已被浏览过或已推送过的媒体内容,以及将剩下的候选的媒体内容标签对应的 媒体内容按照浏览量从高到低的顺序进行排序。
可见, 图 4、 图 5所示的媒体内容推荐设备可以为用户推荐热门并且符合 用户口味的媒体内容, 并且有效地降低推荐给用户的媒体内容的风格差异,避 免推荐给用户冷门的媒体内容, 从而大大提高了用户体验。
根据本发明的一个实施例,图 1所示的媒体内容推荐方法可以由图 4所示 的媒体内容推荐设备中的各个单元来执行。 例如, 图 1所示的步骤 101至 105 可以由图 4所示的计算单元 401、 第一选择单元 402、 查找单元 403、 第二选 择单元 404、 推荐单元 405来执行。
根据本发明的另一个实施例,图 4所示的媒体内容推荐设备中的各个单元 可以分别或全部合并为一个或若干个另外的单元来构成,或者其中的某个(些) 单元还可以再拆分为功能上更小的多个单元来构成。例如但不限于,计算单元 401可以由第一模块 4011及第二模块 4012来实现, 推荐单元 405可以由第三 模块 4051、 第四模块 4052来实现等, 这可以实现同样的操作, 而不影响本发 明的实施例的技术效果的实现。
本领域普通技术人员可以理解上述实施例的各种方法中的全部或部分步 骤是可以通过如图 1所示的媒体内容推荐方法的程序来指令相关的构成如图 4 中所示的媒体内容推荐设备的硬件来完成,该程序可以存储于一计算机可读存 储介质中,并被至少一个处理器执行。其中,所述存储介质可以包括: 闪存盘、 只读存储器 (Read-Only Memory , ROM ) 、 随机存取器 ( Random Access
说明只是用于帮助理解本发明的方法及其核心思想; 同时,对于本领域的一般 技术人员,依据本发明的思想,在具体实施方式及应用范围上均会有改变之处,
— — 综上所述, 本说明书内容不应理解为对本发明的限制。
Claims
1、 一种媒体内容推荐方法, 其特征在于, 包括:
计算媒体内容库中的媒体内容标签的得分;
按照得分从高到低的顺序选取第一门限值个媒体内容标签作为候选的媒 体内容标签;
针对所述候选的媒体内容标签,从所述媒体内容库中查找出所述候选的媒 体内容标签对应的媒体内容;
针对所述候选的媒体内容标签对应的媒体内容,按照浏览量从高到低的顺 序选取第二门限值个媒体内容作为所述候选的媒体内容标签对应的待推荐媒 体内容;
2、 根据权利要求 1所述的方法, 其特征在于, 计算媒体内容库中的媒体 内容标签的得分包括:
针对媒体内容库中的媒体内容标签,计算所述媒体内容库中含有所述媒体 内容标签的媒体内容总数量与所述媒体内容库中所有媒体内容总数量的比值, 以获得所述媒体内容标签热度; 和与所述媒体内容标签热度的比值, 以获得所述媒体内容标签的得分。
3、 根据权利要求 1所述的方法, 其特征在于, 所述按照得分从高到低的 顺序选取第一门限值个媒体内容标签作为候选的媒体内容标签之前,所述方法 还包括:
将每一个媒体内容标签按照得分从高到低的顺序进行排序。
4、 根据权利要求 1〜3任意一项所述的方法, 其特征在于, 所述针对所述 候选的媒体内容标签对应的媒体内容,按照浏览量从高到低的顺序选取第二门 限值个媒体内容作为所述候选的媒体内容标签对应的待推荐媒体内容之前,所 述方法还包括: 的媒体内容;
将剩下的所述候选的媒体内容标签对应的媒体内容按照浏览量从高到低
的顺序进行排序。
5、 根据权利要求 4所述的方法, 其特征在于, 所述将所述候选的媒体内 容标签对应的待推荐媒体内容推荐给用户包括:
将每一个所述候选的媒体内容标签对应的待推荐媒体内容按照浏览量从 高到低的顺序进行排序;
将排序好的每一个所述候选的媒体内容标签对应的待推荐媒体内容作为 一个整体,再按照每一个所述候选的媒体内容标签的得分从高到低的顺序进行 排序后推荐给用户。
6、 根据权利要求 1〜3任意一项所述的方法, 其特征在于, 所述媒体内容 库中的媒体内容包括歌曲、 视频、 语音、 图片以及文本中的任意一种或几种。
7、 根据权利要求 1〜3任意一项所述的方法, 其特征在于, 所述浏览量是 媒体内容的听众数量或媒体内容的点击数量。
8、 一种媒体内容推荐设备, 其特征在于, 包括:
计算单元, 用于计算媒体内容库中的媒体内容标签的得分;
第一选择单元,用于按照得分从高到低的顺序选取第一门限值个媒体内容 标签作为候选的媒体内容标签;
查找单元, 用于针对所述候选的媒体内容标签,从所述媒体内容库中查找 出所述候选的媒体内容标签对应的媒体内容;
第二选择单元, 用于针对所述候选的媒体内容标签对应的媒体内容,按照 浏览量从高到低的顺序选取第二门限值个媒体内容作为所述候选的媒体内容 标签对应的待推荐媒体内容;
推荐单元 ,用于将所述候选的媒体内容标签对应的待推荐媒体内容推荐给 用户。
9、 根据权利要求 8所述的设备, 其特征在于, 所述计算单元包括: 第一模块, 用于针对媒体内容库中的媒体内容标签,计算所述媒体内容库 中含有所述媒体内容标签的媒体内容总数量与所述媒体内容库中所有媒体内 容总数量的比值, 以获得所述媒体内容标签热度; 的用户打分总和与所述媒体内容标签热度的比值,以获得所述媒体内容标签的 得分。
10、 根据权利要求 8所述的设备, 其特征在于,
所述第一选择单元,还用于在按照得分从高到低的顺序选取第一门限值个 媒体内容标签作为候选的媒体内容标签之前 ,将每一个媒体内容标签按照得分 从高到低的顺序进行排序。
11、 根据权利要求 8〜10任意一项所述的设备, 其特征在于,
所述第二选择单元,还用于在针对所述候选的媒体内容标签对应的媒体内 容,按照浏览量从高到低的顺序选取第二门限值个媒体内容作为所述候选的媒 体内容标签对应的待推荐媒体内容之前,滤除所述候选的媒体内容标签对应的 媒体内容中已被浏览过或已推送过的媒体内容,以及将剩下的所述候选的媒体 内容标签对应的媒体内容按照浏览量从高到低的顺序进行排序。
12、 根据权利要求 11所述的设备, 其特征在于, 所述推荐单元包括: 第三模块 ,用于将每一个所述候选的媒体内容标签对应的待推荐媒体内容 按照浏览量从高到低的顺序进行排序;
第四模块 ,用于将排序好的每一个所述候选的媒体内容标签对应的待推荐 媒体内容作为一个整体,再按照每一个所述候选的媒体内容标签的得分从高到 低的顺序进行排序后推荐给用户。
13、 根据权利要求 8〜12任意一项所述的设备, 其特征在于, 所述媒体内 容库中的媒体内容包括歌曲、视频、语音、图片以及文本中的任意一种或几种。
14、 根据权利要求 8所述的设备, 其特征在于, 所述浏览量是媒体内容的 听众数量或媒体内容的点击数量。
15、一种包括程序代码的计算机程序, 当所述计算机程序运行在计算机上 时, 所述程序代码执行根据权利要求 1所述的媒体内容推荐方法的各步骤。
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US14/589,086 US10346412B2 (en) | 2012-08-28 | 2015-01-05 | Method and device for recommendation of media content |
| US16/421,213 US11030202B2 (en) | 2012-08-28 | 2019-05-23 | Method and device for recommendation of media content |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210309267.X | 2012-08-28 | ||
| CN201210309267.XA CN103631823B (zh) | 2012-08-28 | 2012-08-28 | 一种媒体内容推荐方法及设备 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US14/589,086 Continuation US10346412B2 (en) | 2012-08-28 | 2015-01-05 | Method and device for recommendation of media content |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2014032492A1 true WO2014032492A1 (zh) | 2014-03-06 |
Family
ID=50182461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2013/080401 Ceased WO2014032492A1 (zh) | 2012-08-28 | 2013-07-30 | 媒体内容推荐方法及设备 |
Country Status (3)
| Country | Link |
|---|---|
| US (2) | US10346412B2 (zh) |
| CN (1) | CN103631823B (zh) |
| WO (1) | WO2014032492A1 (zh) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3322194A4 (en) * | 2015-07-06 | 2018-05-30 | Tencent Technology (Shenzhen) Company Limited | Video recommendation method, server and storage medium |
| CN108228598A (zh) * | 2016-12-14 | 2018-06-29 | 腾讯科技(北京)有限公司 | 媒体信息排序方法、服务器和系统 |
| CN108595493A (zh) * | 2018-03-15 | 2018-09-28 | 腾讯科技(深圳)有限公司 | 媒体内容的推送方法和装置、存储介质、电子装置 |
| CN110059221A (zh) * | 2019-03-11 | 2019-07-26 | 咪咕视讯科技有限公司 | 视频推荐方法、电子设备及计算机可读存储介质 |
| CN110727813A (zh) * | 2019-08-27 | 2020-01-24 | 达而观信息科技(上海)有限公司 | 一种商品图片的自适应热门指数排序方法 |
| CN114438711A (zh) * | 2020-11-02 | 2022-05-06 | 青岛海尔洗衣机有限公司 | 一种推荐衣物处理设备控制的方法、系统及装置 |
Families Citing this family (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6080649B2 (ja) * | 2013-03-29 | 2017-02-15 | キヤノン株式会社 | レコメンド装置、レコメンド方法及びプログラム |
| CN104133906B (zh) * | 2014-08-06 | 2018-07-31 | 深圳市英威诺科技有限公司 | 一种资讯过滤并智能排序的技术方法 |
| WO2016082164A1 (zh) * | 2014-11-27 | 2016-06-02 | 刘一佳 | 一种游戏的推荐方法以及推荐游戏的装置 |
| CN104462482A (zh) * | 2014-12-18 | 2015-03-25 | 百度在线网络技术(北京)有限公司 | 关于媒体展现的内容提供方法和系统 |
| US10621183B1 (en) * | 2015-03-09 | 2020-04-14 | Interos Solutions, Inc. | Method and system of an opinion search engine with an application programming interface for providing an opinion web portal |
| CN105447087B (zh) * | 2015-11-06 | 2021-08-24 | 腾讯科技(深圳)有限公司 | 一种视频推荐方法及装置 |
| CN105550277A (zh) * | 2015-12-10 | 2016-05-04 | 中国传媒大学 | 基于标签热度的智能电影排名和评价系统 |
| CN107038173B (zh) * | 2016-02-04 | 2021-06-25 | 腾讯科技(深圳)有限公司 | 应用查询方法和装置、相似应用检测方法和装置 |
| US11146865B2 (en) | 2016-03-03 | 2021-10-12 | Comcast Cable Communications, Llc | Determining points of interest in a content item |
| CN107368488A (zh) | 2016-05-12 | 2017-11-21 | 阿里巴巴集团控股有限公司 | 一种确定用户行为偏好的方法、推荐信息的展示方法和装置 |
| CN106168958B (zh) * | 2016-06-28 | 2018-09-07 | 腾讯科技(深圳)有限公司 | 一种音频信息的推荐方法以及服务器 |
| US10223359B2 (en) | 2016-10-10 | 2019-03-05 | The Directv Group, Inc. | Determining recommended media programming from sparse consumption data |
| CN106713985B (zh) * | 2016-12-27 | 2020-07-24 | 广州酷狗计算机科技有限公司 | 一种推荐网络视频的方法和装置 |
| CN107688587B (zh) * | 2017-02-15 | 2023-02-10 | 腾讯科技(深圳)有限公司 | 一种媒体信息展示方法及装置 |
| US11403337B2 (en) * | 2017-12-05 | 2022-08-02 | Google Llc | Identifying videos with inappropriate content by processing search logs |
| CN108055349B (zh) * | 2017-12-28 | 2020-11-13 | 广州酷狗计算机科技有限公司 | 推荐k歌音频的方法、装置和系统 |
| CN109547808B (zh) * | 2018-11-22 | 2021-04-20 | 北京达佳互联信息技术有限公司 | 数据处理方法、装置、服务器及存储介质 |
| CN110222232A (zh) * | 2019-05-10 | 2019-09-10 | 北京百度网讯科技有限公司 | 视频内容处理方法、装置、计算机设备及存储介质 |
| CN110309361B (zh) * | 2019-06-19 | 2021-08-20 | 北京奇艺世纪科技有限公司 | 一种视频评分的确定方法、推荐方法、装置及电子设备 |
| CN110737859B (zh) * | 2019-09-09 | 2022-09-27 | 苏宁云计算有限公司 | 一种up主匹配方法及装置 |
| CN113746874B (zh) * | 2020-05-27 | 2024-04-05 | 百度在线网络技术(北京)有限公司 | 一种语音包推荐方法、装置、设备及存储介质 |
| CN111782957B (zh) * | 2020-07-13 | 2025-07-29 | 腾讯科技(深圳)有限公司 | 一种推荐内容的方法、装置、计算设备及存储介质 |
| US12231731B2 (en) * | 2021-03-30 | 2025-02-18 | Boe Technology Group Co., Ltd. | Method and device for recommending real-time audios and/or videos, and computer storage medium |
| CN115687680A (zh) * | 2021-07-26 | 2023-02-03 | 脸萌有限公司 | 音乐筛选方法、装置、设备、存储介质及程序产品 |
| US11451870B1 (en) | 2021-08-19 | 2022-09-20 | Rovi Guides, Inc. | Methods and systems to dynamically adjust a playlist based on cumulative mood score |
| CN113656697B (zh) * | 2021-08-24 | 2023-12-12 | 北京字跳网络技术有限公司 | 对象推荐方法、装置、电子设备、存储介质 |
| CN114547455B (zh) * | 2022-02-22 | 2024-04-19 | 北京金堤科技有限公司 | 热门对象确定方法、装置、存储介质及电子设备 |
| CN115291503B (zh) * | 2022-09-29 | 2023-03-24 | 北京达佳互联信息技术有限公司 | 推荐方法、推荐装置、电子设备和存储介质 |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101984437A (zh) * | 2010-11-23 | 2011-03-09 | 亿览在线网络技术(北京)有限公司 | 音乐资源个性化推荐方法及系统 |
| CN102043781A (zh) * | 2009-10-23 | 2011-05-04 | 华为技术有限公司 | 一种网页资源推荐方法及装置 |
| 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 |
Family Cites Families (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6598054B2 (en) * | 1999-01-26 | 2003-07-22 | Xerox Corporation | System and method for clustering data objects in a collection |
| US6963867B2 (en) * | 1999-12-08 | 2005-11-08 | A9.Com, Inc. | Search query processing to provide category-ranked presentation of search results |
| US7689615B2 (en) * | 2005-02-25 | 2010-03-30 | Microsoft Corporation | Ranking results using multiple nested ranking |
| US20080235216A1 (en) * | 2007-03-23 | 2008-09-25 | Ruttenberg Steven E | Method of predicitng affinity between entities |
| US8751507B2 (en) * | 2007-06-29 | 2014-06-10 | Amazon Technologies, Inc. | Recommendation system with multiple integrated recommenders |
| US8375073B1 (en) * | 2007-11-12 | 2013-02-12 | Google Inc. | Identification and ranking of news stories of interest |
| US10152721B2 (en) * | 2007-11-29 | 2018-12-11 | International Business Machines Corporation | Aggregate scoring of tagged content across social bookmarking systems |
| US9311391B2 (en) * | 2008-12-30 | 2016-04-12 | Telecom Italia S.P.A. | Method and system of content recommendation |
| US8412796B2 (en) * | 2009-07-31 | 2013-04-02 | University College Dublin—National University of Ireland, Dublin | Real time information feed processing |
| US8924376B1 (en) * | 2010-01-31 | 2014-12-30 | Bryant Christopher Lee | Method for human ranking of search results |
| CN103384883B (zh) * | 2010-06-03 | 2016-11-09 | 汤姆森特许公司 | 利用Top-K处理使语义丰富 |
| US8887058B2 (en) * | 2010-10-25 | 2014-11-11 | Warner Bros. Entertainment Inc. | Media management for multi-user group |
| US9311308B2 (en) * | 2010-10-29 | 2016-04-12 | Hewlett-Packard Development Company, L.P. | Content recommendation for groups |
| US8676742B2 (en) * | 2011-03-29 | 2014-03-18 | Manyworlds, Inc. | Contextual scope-based discovery systems |
| US20130166350A1 (en) * | 2011-06-28 | 2013-06-27 | Smart Software, Inc. | Cluster based processing for forecasting intermittent demand |
| US8805751B2 (en) * | 2011-10-13 | 2014-08-12 | Verizon Patent And Licensing Inc. | User class based media content recommendation methods and systems |
| US8370348B1 (en) * | 2011-12-06 | 2013-02-05 | Google Inc. | Magazine edition recommendations |
| US20130275416A1 (en) * | 2012-04-11 | 2013-10-17 | Avaya Inc. | Scoring of resource groups |
| US9619520B2 (en) * | 2012-06-29 | 2017-04-11 | Papalove Productions, Llc | Method and system for evaluating and sharing media |
| US10055411B2 (en) * | 2015-10-30 | 2018-08-21 | International Business Machines Corporation | Music recommendation engine |
-
2012
- 2012-08-28 CN CN201210309267.XA patent/CN103631823B/zh active Active
-
2013
- 2013-07-30 WO PCT/CN2013/080401 patent/WO2014032492A1/zh not_active Ceased
-
2015
- 2015-01-05 US US14/589,086 patent/US10346412B2/en active Active
-
2019
- 2019-05-23 US US16/421,213 patent/US11030202B2/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102043781A (zh) * | 2009-10-23 | 2011-05-04 | 华为技术有限公司 | 一种网页资源推荐方法及装置 |
| 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 |
| CN101984437A (zh) * | 2010-11-23 | 2011-03-09 | 亿览在线网络技术(北京)有限公司 | 音乐资源个性化推荐方法及系统 |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3322194A4 (en) * | 2015-07-06 | 2018-05-30 | Tencent Technology (Shenzhen) Company Limited | Video recommendation method, server and storage medium |
| CN108228598A (zh) * | 2016-12-14 | 2018-06-29 | 腾讯科技(北京)有限公司 | 媒体信息排序方法、服务器和系统 |
| CN108595493A (zh) * | 2018-03-15 | 2018-09-28 | 腾讯科技(深圳)有限公司 | 媒体内容的推送方法和装置、存储介质、电子装置 |
| CN110059221A (zh) * | 2019-03-11 | 2019-07-26 | 咪咕视讯科技有限公司 | 视频推荐方法、电子设备及计算机可读存储介质 |
| CN110059221B (zh) * | 2019-03-11 | 2023-10-20 | 咪咕视讯科技有限公司 | 视频推荐方法、电子设备及计算机可读存储介质 |
| CN110727813A (zh) * | 2019-08-27 | 2020-01-24 | 达而观信息科技(上海)有限公司 | 一种商品图片的自适应热门指数排序方法 |
| CN110727813B (zh) * | 2019-08-27 | 2023-06-09 | 达而观信息科技(上海)有限公司 | 一种商品图片的自适应热门指数排序方法 |
| CN114438711A (zh) * | 2020-11-02 | 2022-05-06 | 青岛海尔洗衣机有限公司 | 一种推荐衣物处理设备控制的方法、系统及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103631823B (zh) | 2017-01-18 |
| US10346412B2 (en) | 2019-07-09 |
| CN103631823A (zh) | 2014-03-12 |
| US11030202B2 (en) | 2021-06-08 |
| US20160004699A1 (en) | 2016-01-07 |
| US20190278778A1 (en) | 2019-09-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2014032492A1 (zh) | 媒体内容推荐方法及设备 | |
| US11151145B2 (en) | Tag selection and recommendation to a user of a content hosting service | |
| US10853415B2 (en) | Systems and methods of classifying content items | |
| US10579667B2 (en) | Software preference affinity recommendation systems and methods | |
| US9165255B1 (en) | Automatic sequencing of video playlists based on mood classification of each video and video cluster transitions | |
| KR101994987B1 (ko) | 관련 엔티티들 | |
| US20140317105A1 (en) | Live recommendation generation | |
| US9615136B1 (en) | Video classification | |
| CN104008138B (zh) | 一种基于社交网络的音乐推荐方法 | |
| KR102111082B1 (ko) | 지식 패널들을 컨텍스트화 | |
| US20170161818A1 (en) | Explanations for personalized recommendations | |
| Deldjoo et al. | MMTF-14K: a multifaceted movie trailer feature dataset for recommendation and retrieval | |
| US20130166649A1 (en) | System for social-based filtering and contributed playlist generation | |
| US20130332842A1 (en) | Systems and Methods of Selecting Content Items | |
| Ali et al. | Movie recommendation system using genome tags and content-based filtering | |
| WO2015054361A1 (en) | Systems, methods, and computer program products for providing contextually-aware video recommendation | |
| WO2017096877A1 (zh) | 一种推荐方法和装置 | |
| US10262060B1 (en) | Systems and methods for facilitating searching, labeling, and/or filtering of digital media items | |
| US8489571B2 (en) | Digital resources searching and mining through collaborative judgment and dynamic index evolution | |
| CN102129444A (zh) | 信息处理装置、信息处理方法和程序 | |
| WO2011101527A1 (en) | Method for providing a recommendation to a user | |
| EP3144825A1 (en) | Enhanced digital media indexing and retrieval | |
| US20170075999A1 (en) | Enhanced digital media indexing and retrieval | |
| WO2013097078A1 (zh) | 一种视频搜索方法及视频搜索系统 | |
| CN114356979B (zh) | 一种查询方法及其相关设备 |
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: 13832988 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205N DATED 07/05/2015) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 13832988 Country of ref document: EP Kind code of ref document: A1 |
