CN106446052A - Video-on-demand program recommendation method based on user set - Google Patents
Video-on-demand program recommendation method based on user set Download PDFInfo
- Publication number
- CN106446052A CN106446052A CN201610798215.1A CN201610798215A CN106446052A CN 106446052 A CN106446052 A CN 106446052A CN 201610798215 A CN201610798215 A CN 201610798215A CN 106446052 A CN106446052 A CN 106446052A
- Authority
- CN
- China
- Prior art keywords
- user
- film
- recommended
- type
- recommendation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—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
Abstract
The invention discloses a video-on-demand program recommendation method based on a user set. The method comprises the following steps of: 1) collecting the watching records of mass users, setting basic user types, matching the basic user types with film classification labels, and classifying a label weight table for different basic user types; 2) collecting the watching records of the users, carrying out calculation to obtain the model FUser (Li) of a user to be recommended, determining the type of the user to be recommended, and taking a corresponding film classification label weight table as a filtering rule; and 3) collecting the watching records of the users, using an ALS (Alternating Least Squares) algorithm to generate a film recommendation list, and generating a final film recommendation list through the filtering rule. By use of the method, the accuracy and the coverage rate of video recommendation oriented to television end users are high and are unlikely to be affected by popular videos, and a phenomenon that video recommendation information received by the user is repeated is avoided. By use of the method, on the basis of the ALS recommendation algorithm, the characteristics of the television end users are combined to optimize a recommendation result and improve the watching experience of the television end users.
Description
Technical field
A kind of the present invention relates to internet video polymerization technique field, it particularly relates to video based on user's set
Method is recommended in request program.
Background technology
With the development of Internet technology, how massive video being pushed to user becomes problem demanding prompt solution, and
Existing television internet video proposed algorithm be all based on greatly the collaborative filtering based on buyer that traditional electric business website proposes or
Collaborative filtering based on commodity.Both approaches are relatively easy, realize easily, but recommend accuracy and coverage rate not to be very
Preferable.
Currently, the proposed algorithm of video field is mostly derived from the collaborative filtering method (association based on user based on buyer
Same Filtration), its method characteristic is to calculate video similarity matrix (this matrix energy according to the record that mass users watch video
The reflection similarity between video two-by-two), then the viewing record according to unique user, selects the video seen with user respectively
Several similar videos.Consequently recommended video is according to several videos before most like after sequencing of similarity.
But this algorithm is highly susceptible to the impact of popular video, the recommendation video that user receives is often that most people is seen
The popular video crossed.The video having seen user therefore may be led to recommend user again, be additionally, since the film heat of recommendation
Degree is very high to also result in the low problem of valut film recommendation coverage rate.It is true that there being a lot of algorithms to add for temperature problem
The correction of penalty factor, but the principle of this algorithm also results in another question, that is, after recommending a period of time, each
The content recommendation that user receives duplicates.Trace it to its cause be the method be come fuzzy inference user characteristicses by user behavior, not
Using real user feature, particularly television user characteristicses.
For the problem in correlation technique, effective solution is not yet proposed at present.
Content of the invention
For the above-mentioned technical problem in correlation technique, the present invention proposes a kind of video frequency request program based on user's set
Recommendation method, for solving above-mentioned technical problem.
For realizing above-mentioned technical purpose, the technical scheme is that and be achieved in that：
A kind of video frequency request program based on user's set recommends method, including：
S1 collects the viewing record of mass users, and setting elemental user type simultaneously mates film tag along sort, for different basic use
Family type respectively arranges a film tag along sort weight table；
S2 collects the viewing record of mass users and the viewing record of user to be recommended, calculates the mould of user to be recommended
Type FUser(Li), determine the type of user to be recommended and using corresponding film tag along sort weight table as filter rule；
S3 collects the viewing record of mass users and the viewing record of user to be recommended, generates film recommendation using ALS algorithm
Film recommendation list is resequenced by list by filter rule, generates final film recommendation list.
Further, step S1 includes：
S11 collects the viewing record of mass users, and user is divided into 5 elemental user types, is each type of user's coupling
Film tag along sort；
The film tag along sort that S12 is comprised according to each type user, determines the weighted value of each film tag along sort, finally
Respectively generate a film tag along sort weight table for 5 elemental user types.
Further, step S2 includes：
S21 collects the viewing record of mass users and the viewing record of user to be recommended, calculates user's to be recommended
Model FUser(Li)；
S22 is according to the model F of user to be recommendedUser(Li), determine the type of user to be recommended；
If the type of S23 user to be recommended belongs to 5 elemental user types, by corresponding for this user type film contingency table
Sign weight table as filter rule；
If the type of S24 user to be recommended is not belonging to 5 elemental user types, calculate two pressing close to most with it basic
User type, is weighted obtaining new film tag along sort to the film tag along sort weight table of this two elemental user types
Weight table is as filter rule.
Further, in step s 24, if the type of user to be recommended is not belonging to 5 elemental user types, basis
COS distance calculates the two elemental user types pressed close to most with it.
Further, step S3 includes：
S31 collects the viewing record of mass users and the viewing record of user to be recommended, generates film recommendation using ALS algorithm
List；
Film in film recommendation list is carried out descending according to the classified weight value in filtering rule by S32, generates final
Film recommendation list.
Further, in step S31, generate film recommendation list using ALS algorithm and specifically include：
S311 collects the record that user watches film, and according to the record that user watches film, film is given a mark；
S312, according to the score of each film, generates film recommendation list.
Further, in step S311, the record that user watches film refers to film title and the viewing that user watches
The duration of film.
Beneficial effects of the present invention：This method is in the accuracy rate of the video recommendations towards TV end subscriber and coverage rate very
High it is not easy to be subject to popular video to be affected, it is to avoid the phenomenon of user's received video recommendations information repetition；This method exists
ALS combines the feature of TV end subscriber on the basis of recommending to calculate, optimize recommendation results, improve the viewing of TV end subscriber
Experience.
Brief description
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to institute in embodiment
Need use accompanying drawing be briefly described it should be apparent that, drawings in the following description be only the present invention some enforcement
Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also obtain according to these accompanying drawings
Obtain other accompanying drawings.
A kind of video frequency request program based on user's set that Fig. 1 is described according to embodiments of the present invention recommends the stream of method
Journey block diagram.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation description is it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, the every other embodiment that those of ordinary skill in the art are obtained, broadly fall into present invention protection
Scope.
As shown in figure 1, a kind of video frequency request program recommendation side based on user's set described according to embodiments of the present invention
Method, including：
S1 collects the viewing record of mass users, and setting elemental user type simultaneously mates film tag along sort, for different basic use
Family type respectively arranges a film tag along sort weight table；
S2 collects the viewing record of mass users and the viewing record of user to be recommended, calculates the mould of user to be recommended
Type FUser(Li), determine the type of user to be recommended and using corresponding film tag along sort weight table as filter rule；
S3 collects the viewing record of mass users and the viewing record of user to be recommended, generates film recommendation using ALS algorithm
Film recommendation list is resequenced by list by filter rule, generates final film recommendation list.
Wherein, step S1 further includes：
S11 collects the viewing record of mass users, and user is divided into 5 elemental user types, is each type of user's coupling
Film tag along sort；
The film tag along sort that S12 is comprised according to each type user, determines the weighted value of each film tag along sort, finally
Respectively generate a film tag along sort weight table for 5 elemental user types.
Wherein, step S2 further includes：
S21 collects the viewing record of mass users and the viewing record of user to be recommended, calculates user's to be recommended
Model FUser(Li)；
S22 is according to the model F of user to be recommendedUser(Li), determine the type of user to be recommended；
If the type of S23 user to be recommended belongs to 5 elemental user types, by corresponding for this user type film contingency table
Sign weight table as filter rule；
If the type of S24 user to be recommended is not belonging to 5 elemental user types, calculate two pressing close to most with it basic
User type, is weighted obtaining new film tag along sort to the film tag along sort weight table of this two elemental user types
Weight table is as filter rule.
Wherein, in step s 24, if the type of user to be recommended is not belonging to 5 elemental user types, according to cosine
Distance calculates the two elemental user types pressed close to most with it.
Wherein, step S3 further includes：
S31 collects the viewing record of mass users and the viewing record of user to be recommended, generates film recommendation using ALS algorithm
List；
Film in film recommendation list is carried out descending according to the classified weight value in filtering rule by S32, generates final
Film recommendation list.
Wherein, in step S31, generate film recommendation list using ALS algorithm and specifically include：
S311 collects the record that user watches film, and according to the record that user watches film, film is given a mark；
S312, according to the score of each film, generates film recommendation list.
Wherein, in step S311, the record that user watches film refers to the film title of user's viewing and watches film
Duration.
Understand the technique scheme of the present invention for convenience, below by way of above-mentioned to the present invention in specifically used mode
Technical scheme is described in detail.
It is assumed that the user model of calculated user to be recommended is F when specifically usedUser=(0.3,0.2,0.1,
1,0), then one kind may be just to have had child for this family, and youngster oneself carries small children the family of (not having old man).And this user
Exactly one of 5 kinds of fundamental type users.
It is calculated the corresponding film classified weight table of elemental user type belonging to this user as shown in table 1, mistake
Filter rule also as shown in table 1 (if user is not belonging to 5 kinds of fundamental types, the form of the filter rule ultimately generating also with table 1 one
Cause).
Table 1：
Finally it is assumed that the preliminary recommendation list being obtained by ALS algorithm is as shown in table 2, the output of wherein ALS algorithm only has
Film title, tag along sort information comes from back-end data, the label more than one of each film.
Table 2：
Consequently recommended list after resequencing after employing filter rule may be for shown in table 3.
Table 3：
In sum, by means of the technique scheme of the present invention, this method is in the video recommendations towards TV end subscriber
Accuracy rate and coverage rate very high it is not easy to affected by popular video, it is to avoid user's received video recommendations information
The phenomenon repeating；This method combines the feature of TV end subscriber on the basis of ALS recommends to calculate, and optimizes recommendation results, carries
The high viewing experience of TV end subscriber.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all essences in the present invention
Within god and principle, any modification, equivalent substitution and improvement made etc., should be included within the scope of the present invention.
Claims (7)
1. a kind of video frequency request program based on user's set recommends method it is characterised in that including：
S1 collects the viewing record of mass users, and setting elemental user type simultaneously mates film tag along sort, for different basic use
Family type respectively arranges a film tag along sort weight table；
S2 collects the viewing record of mass users and the viewing record of user to be recommended, calculates the mould of user to be recommended
Type FUser（Li）, determine the type of user to be recommended and using corresponding film tag along sort weight table as filter rule；
S3 collects the viewing record of mass users and the viewing record of user to be recommended, generates film recommendation using ALS algorithm
Film recommendation list is resequenced by list by filter rule, generates final film recommendation list.
2. video frequency request program based on user's set according to claim 1 recommends method it is characterised in that step S1
Further include：
S11 collects the viewing record of mass users, and user is divided into 5 elemental user types, is each type of user's coupling
Film tag along sort；
The film tag along sort that S12 is comprised according to each type user, determines the weighted value of each film tag along sort, finally
Respectively generate a film tag along sort weight table for 5 elemental user types.
3. video frequency request program based on user's set according to claim 2 recommends method it is characterised in that step S2
Further include：
S21 collects the viewing record of mass users and the viewing record of user to be recommended, calculates user's to be recommended
Model FUser（Li）；
S22 is according to the model F of user to be recommendedUser（Li）, determine the type of user to be recommended；
If the type of S23 user to be recommended belongs to 5 elemental user types, by corresponding for this user type film contingency table
Sign weight table as filter rule；
If the type of S24 user to be recommended is not belonging to 5 elemental user types, calculate two pressing close to most with it basic
User type, is weighted obtaining new film tag along sort to the film tag along sort weight table of this two elemental user types
Weight table is as filter rule.
4. video frequency request program based on user's set according to claim 3 recommends method it is characterised in that in step
In S24, if the type of user to be recommended is not belonging to 5 elemental user types, is calculated according to COS distance and paste most with it
Two near elemental user types.
5. the video frequency request program recommendation method based on user's set according to claim 3 or 4 is it is characterised in that walk
Rapid S3 further includes：
S31 collects the viewing record of mass users and the viewing record of user to be recommended, generates film using ALS algorithm and pushes away
Recommend list；
Film in film recommendation list is carried out descending according to the classified weight value in filtering rule by S32, generates final
Film recommendation list.
6. video frequency request program based on user's set according to claim 5 recommends method it is characterised in that in step
In S31, generate film recommendation list using ALS algorithm and specifically include：
S311 collects the record that user watches film, and according to the record that user watches film, film is given a mark；
S312, according to the score of each film, generates film recommendation list.
7. video frequency request program based on user's set according to claim 6 recommends method it is characterised in that in step
In S311, the record that user watches film refers to the film title of user's viewing and the duration of viewing film.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610798215.1A CN106446052A (en) | 2016-08-31 | 2016-08-31 | Video-on-demand program recommendation method based on user set |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610798215.1A CN106446052A (en) | 2016-08-31 | 2016-08-31 | Video-on-demand program recommendation method based on user set |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106446052A true CN106446052A (en) | 2017-02-22 |
Family
ID=58164544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610798215.1A Pending CN106446052A (en) | 2016-08-31 | 2016-08-31 | Video-on-demand program recommendation method based on user set |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106446052A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107105318A (en) * | 2017-03-21 | 2017-08-29 | 华为技术有限公司 | A kind of video hotspot fragment extracting method, user equipment and server |
| CN110087103A (en) * | 2019-04-26 | 2019-08-02 | 北京奇艺世纪科技有限公司 | A kind of video recommendation system, method, apparatus and computer |
| CN110737833A (en) * | 2019-09-29 | 2020-01-31 | 海南省火蓝数据有限公司 | commodity pushing method based on recommendation degree |
| CN111026977A (en) * | 2019-12-17 | 2020-04-17 | 腾讯科技（深圳）有限公司 | Information recommendation method and device and storage medium |
| CN111028572A (en) * | 2019-12-31 | 2020-04-17 | 浙江正元智慧科技股份有限公司 | Online education platform |
| CN112015736A (en) * | 2020-08-21 | 2020-12-01 | 广州欢网科技有限责任公司 | Spark Mllib-based multifunctional recommendation method and device |
| WO2021042826A1 (en) * | 2019-09-05 | 2021-03-11 | 苏宁云计算有限公司 | Video playback completeness prediction method and apparatus |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102819529A (en) * | 2011-06-10 | 2012-12-12 | 阿里巴巴集团控股有限公司 | Information publishing method and system for social website |
| CN104050258A (en) * | 2014-06-15 | 2014-09-17 | 中国传媒大学 | Group recommendation method based on interest groups |
| US20150088684A1 (en) * | 2013-09-24 | 2015-03-26 | Valerie Nygaard | Social media-based recommendations |
-
2016
- 2016-08-31 CN CN201610798215.1A patent/CN106446052A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102819529A (en) * | 2011-06-10 | 2012-12-12 | 阿里巴巴集团控股有限公司 | Information publishing method and system for social website |
| US20150088684A1 (en) * | 2013-09-24 | 2015-03-26 | Valerie Nygaard | Social media-based recommendations |
| CN104050258A (en) * | 2014-06-15 | 2014-09-17 | 中国传媒大学 | Group recommendation method based on interest groups |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107105318A (en) * | 2017-03-21 | 2017-08-29 | 华为技术有限公司 | A kind of video hotspot fragment extracting method, user equipment and server |
| US11265624B2 (en) | 2017-03-21 | 2022-03-01 | Huawei Technologies Co., Ltd. | Hot video clip extraction method, user equipment, and server |
| CN110087103A (en) * | 2019-04-26 | 2019-08-02 | 北京奇艺世纪科技有限公司 | A kind of video recommendation system, method, apparatus and computer |
| WO2021042826A1 (en) * | 2019-09-05 | 2021-03-11 | 苏宁云计算有限公司 | Video playback completeness prediction method and apparatus |
| CN110737833A (en) * | 2019-09-29 | 2020-01-31 | 海南省火蓝数据有限公司 | commodity pushing method based on recommendation degree |
| CN110737833B (en) * | 2019-09-29 | 2022-09-13 | 安徽火蓝数据有限公司 | Commodity pushing method based on recommendation degree |
| CN111026977A (en) * | 2019-12-17 | 2020-04-17 | 腾讯科技（深圳）有限公司 | Information recommendation method and device and storage medium |
| CN111028572A (en) * | 2019-12-31 | 2020-04-17 | 浙江正元智慧科技股份有限公司 | Online education platform |
| CN112015736A (en) * | 2020-08-21 | 2020-12-01 | 广州欢网科技有限责任公司 | Spark Mllib-based multifunctional recommendation method and device |
| CN112015736B (en) * | 2020-08-21 | 2024-04-05 | 广州欢网科技有限责任公司 | Multi-functional recommendation method and device based on Spark Mllib |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106446052A (en) | Video-on-demand program recommendation method based on user set | |
| Li et al. | Neural attentive session-based recommendation | |
| CN108197330B (en) | Data digging method and device based on social platform | |
| CN103559758B (en) | A kind of intelligentized vending system and good selling method | |
| CN203773629U (en) | Intelligent self-service vending system | |
| CN101482884A (en) | Cooperation recommending system based on user predilection grade distribution | |
| US20150193851A1 (en) | Method, server, client terminal, and electronic commerce system for product comparison | |
| CN105488233A (en) | Reading information recommendation method and system | |
| CN105975472A (en) | Method and device for recommendation | |
| CN105589905A (en) | User interest data analysis and collection system and method | |
| CN103559206A (en) | Information recommending method and system | |
| CN103970906A (en) | Method and device for establishing video tags and method and device for displaying video contents | |
| CN101287082A (en) | Collaborative filtered recommendation method introducing hotness degree weight of program | |
| CN103995839A (en) | Commodity recommendation optimizing method and system based on collaborative filtering | |
| CN102063433A (en) | Method and device for recommending related items | |
| CN102999507A (en) | Recommendation processing method and device for information of network microblog celebrities | |
| CN103106600A (en) | Commodity information push system and commodity information push method | |
| CN106168980A (en) | Multimedia resource recommends sort method and device | |
| CN104298787A (en) | Individual recommendation method and device based on fusion strategy | |
| CN108182268B (en) | Collaborative filtering recommendation method and system based on social network | |
| CN104867015B (en) | A kind of article delivery person recommendation method based on user's moving projection | |
| CN106127541A (en) | A kind of credit assessment method based on negative investigation and system | |
| CN111475739A (en) | Heterogeneous social network user anchor link identification method based on meta-path | |
| CN108833488A (en) | A kind of method for pushing and server of recommendation information | |
| CN103678548B (en) | Failure service based on integrated mode substitutes recommendation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170222 |
|
| RJ01 | Rejection of invention patent application after publication |