CN102306178A - Video recommendation method and device - Google Patents

Video recommendation method and device Download PDF

Info

Publication number
CN102306178A
CN102306178A CN201110247663A CN201110247663A CN102306178A CN 102306178 A CN102306178 A CN 102306178A CN 201110247663 A CN201110247663 A CN 201110247663A CN 201110247663 A CN201110247663 A CN 201110247663A CN 102306178 A CN102306178 A CN 102306178A
Authority
CN
China
Prior art keywords
video
transition probability
cookie
probability matrix
user
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
Application number
CN201110247663A
Other languages
Chinese (zh)
Inventor
赵旭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shengle Information Technolpogy Shanghai Co Ltd
Original Assignee
Shengle Information Technolpogy Shanghai Co Ltd
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 Shengle Information Technolpogy Shanghai Co Ltd filed Critical Shengle Information Technolpogy Shanghai Co Ltd
Priority to CN201110247663A priority Critical patent/CN102306178A/en
Publication of CN102306178A publication Critical patent/CN102306178A/en
Pending legal-status Critical Current

Links

Images

Abstract

The invention provides a video recommendation method and a video recommendation device. Data information in cookies of a log database of a user video search system is utilized to be subjected to total probability transfer, and more related information among videos and more recommended videos are acquired, so that the diversity of a video recommendation result range is improved.

Description

Video recommend method and device
Technical field
The present invention relates to the computer data field of storage, relate in particular to a kind of video recommend method and device.
Background technology
The video search website can offer the thousands of video content of user usually and supply user search to browse and watch, and, all can there be every day the huge emerging video of quantity to be searched record and retrieval service is provided by video website.In the face of like this huge quantity of information, how therefrom to find the video information that meets user's needs more, make and search and watch the video of certain specific area simple more easy.
The video recommended technology promptly occurs for solving this type of problem, and so-called video recommended technology is exactly to pass through certain computerized algorithm, in the video of magnanimity, finds out the interested probably video of those users automatically, and it is offered the user.And the interested probably video of those users is to make prediction according to the video that the user is watching now or watching in the past.Be used as the prediction of user interest and recommend the user through the maximally related video of video that finds those and user watching or watched in the past.So the key that video is recommended is how to find and the maximally related video of given video, and recommends the user.
SESSION is meant the time interval that a terminal user and interactive system communicate; This time interval is shorter usually; The information that the method great majority of the similarity between existing calculating video utilize is confined to the SESSION level; Yet the quantity of information through detection SESSION obtains is limited; And on the algorithm that calculates similarity, some simple characteristic informations have often only been utilized; Such as: the classification of video, the length of video, the issuing time of video etc.These information often make that the result who recommends too concentrates in the very little scope, make the variation of recommendation results receive very big restriction.
Summary of the invention
The object of the present invention is to provide a kind of video recommend method and device, recommend, make the variation of video recommendation results scope get a promotion through using more quantity of information to obtain more video.
For addressing the above problem, the present invention provides a kind of video recommend method, comprising:
Data message from the log database of user video search system among each COOKIE of extraction is as training sample;
It is right to calculate in the said training sample transition probability between all COOKIE and VIDEO, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE;
According to said COOKIE to the transition probability matrix of VIDEO and VIDEO to the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO;
Transition probability matrix according between the said VIDEO obtains recommended models, and embeds said user video search system to return recommendation results to the user.
Further, said COOKIE to the computing formula of the transition probability matrix of VIDEO is:
Figure BDA0000086106030000021
Wherein, i, j is natural number.
Further, said VIDEO to the computing formula of the transition probability matrix of COOKIE is:
Figure BDA0000086106030000022
Wherein, i, j is natural number.
Further, through calculate said COOKIE to the transition probability matrix of VIDEO and VIDEO to the product of the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO.
Further, the transition probability matrix between the said VIDEO is pressed row normalization, and will sort left to right, obtain recommended models.
Accordingly, the present invention also provides a kind of video recommendation apparatus, embeds user's video searching system, and said video recommendation apparatus comprises:
The training sample extraction module is used for extracting data message each COOKIE as training sample from the log database of user video search system;
COOKIE and VIDEO co-occurrence computing module, the transition probability that is used for calculating between all COOKIE of said training sample and the VIDEO is right, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE;
The related computing module with VIDEO of VIDEO is used for according to said COOKIE to the transition probability matrix of VIDEO and VIDEO obtaining the transition probability matrix between the VIDEO to the transition probability matrix of COOKIE;
VIDEO recommended models module is used for obtaining recommended models according to the transition probability matrix between the said VIDEO, and embeds said user video search system to return recommendation results to the user.
Compared with prior art; Video recommend method provided by the invention and device; Data message among the COOKIE of the log database through utilizing the user video search system; Carrying out total probability shifts; Obtain the related information between the more VIDEO; Obtain more video and recommend, make the variation of video recommendation results scope get a promotion.
Description of drawings
Fig. 1 is the video recommend method process flow diagram of the embodiment of the invention one;
Fig. 2 is the structural representation of the video recommendation apparatus of the embodiment of the invention two.
Embodiment
Below in conjunction with accompanying drawing and specific embodiment video recommend method and device that the present invention proposes are done further explain.
Embodiment one
As shown in Figure 1, present embodiment provides a kind of video recommend method, comprises step S101 to step S104, describes step S101 in detail to step S104 below in conjunction with concrete data.
Step S101, the data message from the log database of user video search system among each COOKIE of extraction is as training sample.
In the present embodiment, it is following from the log database of user video search system, to extract the COOKIE data:
COOKIE 1：VIDEO 1，VIDEO 2，VIDEO 3
COOKIE 2：VIDEO 1，VIDEO 3，VIDEO 4
COOKIE 3：VIDEO 5，VIDEO 6，VIDEO 7
With these data as training sample.
Step S102, it is right to calculate in the said training sample transition probability between all COOKIE and VIDEO, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE.
In this step, described in the calculation procedure S101 in the training sample transition probability between all COOKIE and the VIDEO right, i.e. P (COOKIE i| VIDEO j) and P (VIDEI j| COOKIE i), obtain the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE.
Wherein, said COOKIE to the computing formula of the transition probability matrix of VIDEO is:
Figure BDA0000086106030000041
Said VIDEO to the computing formula of the transition probability matrix of COOKIE is:
Figure BDA0000086106030000042
Wherein, i, j is natural number, in the present embodiment, i=1,2,3; J=1,2,3,4,5,6,7; In the present embodiment, work as i=1, during j=3, COOKIE 1To VIDEO 3Transition probability be:
P ( COOKIE 1 | VIDEO 3 ) = 1 1 + 1 = 0.5
In like manner, can calculate other COOKIE iWith VIDEO jBetween transition probability, obtain the transition probability matrix A of COOKIE to VIDEO:
A = 0.5 0.5 0 1 0 0 0.5 0.5 0 0 1 0 0 0 1 0 0 1 0 0 1
According to top same method, can calculate the transition probability matrix B of VIDEO equally to COOKIE:
B = 0.3 0.3 0.3 0 0 0 0 0.3 0 0.3 0.3 0 0 0 0 0 0 0 0.3 0.3 0.3
Step S103, according to said COOKIE to the transition probability matrix of VIDEO and VIDEO to the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO.
In this step, calculate said COOKIE to the transition probability matrix of VIDEO and VIDEO to the product of the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO, be formulated as: P (VIDEO i| VIDEO j)=∑ P (VIDEO i| COOKIE K)P (COOKIE k| VIDEO j).
In the present embodiment, the transition probability matrix C between VIDEO and the VIDEO is the product matrix of matrix A and matrix B just, promptly
C = AgB 0.5 0,5 0 1 0 0 0.5 0.5 0 0 1 0 0 0 1 0 0 1 0 0 1 0.3 0.3 0.3 0 0 0 0 0.3 0 0.3 0.3 0 0 0 0 0 0 0 0.3 0.3 0.3 = 0.3 0.15 0.3 0.15 0.15 0.15 0.15 0.3 0.3 0.3 0 0 0 0 0.3 0.15 0.3 0.15 0.15 0.15 0.15 0.3 0 0.3 0.3 0 0 0 0 0 0 0 0.3 0.3 0.3 0 0 0 0 0.3 0.3 0.3 0 0 0 0 0.3 0.3 0.3 ;
Step S104 obtains recommended models according to the transition probability matrix between the said VIDEO, and embeds said user video search system to return recommendation results to the user.
In this step, the transition probability matrix between the said VIDEO by row normalization, is obtained:
Figure BDA0000086106030000053
Product matrix C by after the row normalization is exactly a recommended models, each Elements C IjExpression VIDEO iWith VIDEO jSimilarity.
When recommending, suppose that the user has watched VIDEO 3, C sorts left to right to product matrix, then can in the third line of product matrix C, search the element with greatest measure, and the VIDEO that it is corresponding the most recommendation results return.Here maximum two elements, 0.22 corresponding VIDEO is respectively VIDEO in the third line of product matrix C 1And VIDEO 3, again because VIDEO 3The video of having watched for the user is so only return VIDEO 1As recommendation results.
Embodiment two
As shown in Figure 2, present embodiment two provides a kind of video recommendation apparatus 2, embeds in user's the video searching system 1, and said video recommendation apparatus 2 comprises:
Training sample extraction module 21 is used for extracting data message each COOKIE as training sample from the log database of user video search system;
COOKIE and VIDEO co-occurrence probabilities module 22, the transition probability that is used for calculating between all COOKIE of said training sample and the VIDEO is right, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE;
VIDEO and VIDEO association probability module 23 are used for according to said COOKIE to the transition probability matrix of VIDEO and VIDEO obtaining the transition probability matrix between the VIDEO to the transition probability matrix of COOKIE;
VIDEO recommended models module 24 is used for obtaining recommended models according to the transition probability matrix between the said VIDEO, and embeds said user video search system 1 to return recommendation results to the user.
In sum; Video recommend method provided by the invention and device; Data message among the COOKIE of the log database through utilizing the user video search system; Carrying out total probability shifts; Obtain the related information between the more VIDEO; Obtain more video and recommend, make the variation of video recommendation results scope get a promotion.
Obviously, those skilled in the art can carry out various changes and modification to invention and not break away from the spirit and scope of the present invention.Like this, belong within the scope of claim of the present invention and equivalent technologies thereof if of the present invention these are revised with modification, then the present invention also is intended to comprise these changes and modification interior.

Claims (10)

1. a video recommend method is characterized in that, comprising:
Data message from the log database of user video search system among each COOKIE of extraction is as training sample;
It is right to calculate in the said training sample transition probability between all COOKIE and VIDEO, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE;
According to said COOKIE to the transition probability matrix of VIDEO and VIDEO to the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO;
Transition probability matrix according between the said VIDEO obtains recommended models, and embeds said user video search system to return recommendation results to the user.
2. video recommend method as claimed in claim 1 is characterized in that, said COOKIE to the computing formula of the transition probability matrix of VIDEO is:
Figure FDA0000086106020000011
Wherein, i, j is natural number.
3. video recommend method as claimed in claim 1 is characterized in that, said VIDEO to the computing formula of the transition probability matrix of COOKIE is:
Figure FDA0000086106020000012
Wherein, i, j is natural number.
4. video recommend method as claimed in claim 1 is characterized in that, through calculate said COOKIE to the transition probability matrix of VIDEO and VIDEO to the product of the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO.
5. video recommend method as claimed in claim 1 is characterized in that, the transition probability matrix between the said VIDEO is pressed row normalization, and will sort left to right, and obtains recommended models.
6. video recommendation apparatus embeds user's video searching system, it is characterized in that said video recommendation apparatus comprises:
The training sample extraction module is used for extracting data message each COOKIE as training sample from the log database of user video search system;
COOKIE and VIDEO co-occurrence computing module, the transition probability that is used for calculating between all COOKIE of said training sample and the VIDEO is right, obtains the transition probability matrix of COOKIE to the transition probability matrix of VIDEO and VIDEO to COOKIE;
The related computing module with VIDEO of VIDEO is used for according to said COOKIE to the transition probability matrix of VIDEO and VIDEO obtaining the transition probability matrix between the VIDEO to the transition probability matrix of COOKIE;
VIDEO recommended models module is used for obtaining recommended models according to the transition probability matrix between the said VIDEO, and embeds said user video search system to return recommendation results to the user.
7. video recommendation apparatus as claimed in claim 6 is characterized in that, said COOKIE to the computing formula of the transition probability matrix of VIDEO is:
Figure FDA0000086106020000021
Wherein, i, j is natural number.
8. video recommendation apparatus as claimed in claim 6 is characterized in that, said VIDEO to the computing formula of the transition probability matrix of COOKIE is:
Figure FDA0000086106020000022
Wherein, i, j is natural number.
9. video recommendation apparatus as claimed in claim 6 is characterized in that, through calculate said COOKIE to the transition probability matrix of VIDEO and VIDEO to the product of the transition probability matrix of COOKIE, obtain the transition probability matrix between the VIDEO.
10. video recommendation apparatus as claimed in claim 6 is characterized in that, the transition probability matrix between the said VIDEO is pressed row normalization, and will sort left to right, and obtains recommended models.
CN201110247663A 2011-08-25 2011-08-25 Video recommendation method and device Pending CN102306178A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201110247663A CN102306178A (en) 2011-08-25 2011-08-25 Video recommendation method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201110247663A CN102306178A (en) 2011-08-25 2011-08-25 Video recommendation method and device

Publications (1)

Publication Number Publication Date
CN102306178A true CN102306178A (en) 2012-01-04

Family

ID=45380040

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201110247663A Pending CN102306178A (en) 2011-08-25 2011-08-25 Video recommendation method and device

Country Status (1)

Country Link
CN (1) CN102306178A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103686237A (en) * 2013-11-19 2014-03-26 乐视致新电子科技（天津）有限公司 Method and system for recommending video resource
CN104238516A (en) * 2014-09-15 2014-12-24 厦门大学 Method for monitoring state of boiler system equipment
CN105183925A (en) * 2015-10-30 2015-12-23 合一网络技术(北京)有限公司 Content association recommending method and content association recommending device
CN105302880A (en) * 2015-10-14 2016-02-03 合一网络技术(北京)有限公司 Content correlation recommendation method and apparatus
CN105916034A (en) * 2016-04-14 2016-08-31 乐视控股（北京）有限公司 Video recommendation method and apparatus
CN106547768A (en) * 2015-09-21 2017-03-29 中兴通讯股份有限公司 A kind of control method for playing back and device of media file

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101453361A (en) * 2007-12-07 2009-06-10 中国科学院声学研究所 Website request queue management method
CN101826114A (en) * 2010-05-26 2010-09-08 南京大学 Multi Markov chain-based content recommendation method
CN101923545A (en) * 2009-06-15 2010-12-22 北京百分通联传媒技术有限公司 Method for recommending personalized information

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101453361A (en) * 2007-12-07 2009-06-10 中国科学院声学研究所 Website request queue management method
CN101923545A (en) * 2009-06-15 2010-12-22 北京百分通联传媒技术有限公司 Method for recommending personalized information
CN101826114A (en) * 2010-05-26 2010-09-08 南京大学 Multi Markov chain-based content recommendation method

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103686237A (en) * 2013-11-19 2014-03-26 乐视致新电子科技（天津）有限公司 Method and system for recommending video resource
CN103686237B (en) * 2013-11-19 2017-03-08 乐视致新电子科技（天津）有限公司 Recommend the method and system of video resource
CN104238516A (en) * 2014-09-15 2014-12-24 厦门大学 Method for monitoring state of boiler system equipment
CN106547768A (en) * 2015-09-21 2017-03-29 中兴通讯股份有限公司 A kind of control method for playing back and device of media file
CN105302880A (en) * 2015-10-14 2016-02-03 合一网络技术(北京)有限公司 Content correlation recommendation method and apparatus
CN105183925A (en) * 2015-10-30 2015-12-23 合一网络技术(北京)有限公司 Content association recommending method and content association recommending device
CN105916034A (en) * 2016-04-14 2016-08-31 乐视控股（北京）有限公司 Video recommendation method and apparatus

Similar Documents

Publication Publication Date Title
CN108805598B (en) Similarity information determination method, server and computer-readable storage medium
CN111008321B (en) Logistic regression recommendation-based method, device, computing equipment and readable storage medium
US20070198459A1 (en) System and method for online information analysis
CN102509233A (en) User online action information-based recommendation method
CN106407420B (en) Multimedia resource recommendation method and system
CN111797214A (en) FAQ database-based problem screening method and device, computer equipment and medium
CN112434151A (en) Patent recommendation method and device, computer equipment and storage medium
CN103886067A (en) Method for recommending books through label implied topic
CN102306178A (en) Video recommendation method and device
CN108763321A (en) A kind of related entities recommendation method based on extensive related entities network
CN106062743A (en) Systems and methods for keyword suggestion
CN104615741B (en) Cold-start project recommendation method and device based on cloud computing
US20140297628A1 (en) Text Information Processing Apparatus, Text Information Processing Method, and Computer Usable Medium Having Text Information Processing Program Embodied Therein
CN105468649B (en) Method and device for judging matching of objects to be displayed
CN111815375B (en) User portrayal method and device in advertisement putting
CN105787662A (en) Mobile application software performance prediction method based on attributes
CN114266443A (en) Data evaluation method and device, electronic equipment and storage medium
CN103577472A (en) Method and system for obtaining and presuming personal information as well as method and system for classifying and retrieving commodities
US20220321963A1 (en) Method and apparatus for video recommendation, and refrigerator with screen
CN112989118A (en) Video recall method and device
CN112733006B (en) User portrait generation method, device and equipment and storage medium
CN111831913A (en) Potential competitor information recommendation method, device, equipment and storage medium
CN106446696B (en) Information processing method and electronic equipment
WO2023082864A1 (en) Training method and apparatus for content recommendation model, device, and storage medium
CN103678720B (en) Method and device for processing user feedback data

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20120104