CN109947983A - Video recommendation method, system, terminal and computer readable storage medium - Google Patents
Video recommendation method, system, terminal and computer readable storage medium Download PDFInfo
- Publication number
- CN109947983A CN109947983A CN201710850056.XA CN201710850056A CN109947983A CN 109947983 A CN109947983 A CN 109947983A CN 201710850056 A CN201710850056 A CN 201710850056A CN 109947983 A CN109947983 A CN 109947983A
- Authority
- CN
- China
- Prior art keywords
- matrix
- video
- user
- log data
- singular vector
- 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
Landscapes
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The present invention provides a kind of video recommendation method, system, terminal and computer readable storage mediums, and wherein method includes: to obtain the user behaviors log data of user's history viewing video；User-video rating matrix is constructed according to user behaviors log data；Full-rank factorization is carried out to user-video rating matrix, obtains left singular vector matrix corresponding with user-video matrix, right singular vector matrix and diagonal matrix；The similar neighborhood set of user is calculated according to left singular vector matrix, right singular vector matrix and diagonal matrix；Video recommendations are carried out according to similar neighborhood set and user behaviors log data acquisition video recommendations list, and according to video recommendations list.The present invention can solve the problem of existing Sparse generally existing based on user in collaborative filtering-video rating matrix, promote the video recommendations effect of recommender system；The time complexity that user or video arest neighbors calculate during video recommendations can also be reduced.
Description
Technical field
The invention belongs to video recommendations technical field more particularly to a kind of video recommendation method, system, terminal and computers
Readable storage medium storing program for executing.
Background technique
Universal with mobile device with the rapid development of mobile Internet, video-see has become in people's life and disappears
Time indispensable a part is ground, user's viscosity how is effectively improved and retention ratio is that current major video operator emphasis closes
The problem of note.Video recommendation system can be under the premise of user be without having a definite purpose, and active recommended user may interested view
Frequency information, it is highly effective to raising user viscosity and retention ratio, it is widely used by major video operation side at present.
The core of video recommendation system first is that proposed algorithm, the layout strategy or method choice of proposed algorithm are by direct shadow
Ring entire recommender system whether superiority and inferiority, existing video recommendations algorithm is roughly divided into three classes: socialization proposed algorithm, based on content
Proposed algorithm and Collaborative Filtering Recommendation Algorithm, wherein Collaborative Filtering Recommendation Algorithm only needs the user behaviors log data of user that can produce
Raw good recommendation effect, thus it is widely used in the video recommendation system of major major video website.However, currently based on
There are following problems for the video recommendation system of Collaborative Filtering Recommendation Algorithm:
1) Collaborative Filtering Recommendation Algorithm be all based on greatly user user behaviors log data generate " user-video " rating matrix into
Row video recommendations, if the limited amount of user's history viewing video, " user-video " rating matrix can have serious data
Sparse Problems lack this general character of same subscriber so as to cause lacking between a large number of users, use between same video or multitude of video
The arest neighbors list at family is almost sky, and it is ineffective to eventually lead to video recommendations；
2) dimension of being continuously increased with video user, " user-video " rating matrix will be higher and higher, after causing
The time complexity that user or video arest neighbors calculate in continuous video recommendations calculating process is very high, the video recommendations of recommender system
As a result it possibly can not timely update, influence user experience.
Summary of the invention
In view of this, the present invention provides a kind of video recommendation method, system, terminal and computer readable storage medium,
If to solve the limited amount of user's history viewing video in the prior art, " user-video " rating matrix can exist serious
Sparse, cause video recommendations ineffective and being continuously increased with video user, user during video recommendations
Or the higher problem of time complexity that video arest neighbors calculates.
The first aspect of the present invention provides a kind of video recommendation method, comprising:
Obtain the user behaviors log data of user's history viewing video；
User-video rating matrix is constructed according to the user behaviors log data；
It is completely lost decomposition to the user-video rating matrix, obtains left surprise corresponding with the user-video matrix
Incorgruous moment matrix, right singular vector matrix and diagonal matrix；
It is calculated according to the left singular vector matrix, the right singular vector matrix and the diagonal matrix described
The similar neighborhood set of user；
According to the similar neighborhood set and the user behaviors log data acquisition video recommendations list, and according to the video
Recommendation list carries out video recommendations.
The second aspect of the present invention provides a kind of video recommendation system, comprising:
Behavioral data acquiring unit, for obtaining the user behaviors log data of user's history viewing video；
Rating matrix construction unit, for constructing user-video rating matrix according to the user behaviors log data；
Rating matrix decomposition unit obtains and the use for being completely lost decomposition to the user-video rating matrix
The corresponding left singular vector matrix of family-video matrix, right singular vector matrix and diagonal matrix；
Similar neighborhood computing unit, for according to the left singular vector matrix, the right singular vector matrix and institute
State the similar neighborhood set that the user is calculated in diagonal matrix；
Video recommendations unit, for being arranged according to the similar neighborhood set and the user behaviors log data acquisition video recommendations
Table, and video recommendations are carried out according to the video recommendations list.
The third aspect of the present invention provides a kind of terminal, including memory, processor and is stored in the memory
In and the computer program that can run on the processor, wherein the processor is realized when executing the computer program
Such as the step of above-mentioned first aspect the method.
The fourth aspect of the present invention provides a kind of computer readable storage medium, and the computer readable storage medium is deposited
Contain computer program, wherein realize when the computer program is executed by processor such as above-mentioned first aspect the method
Step.
User behaviors log data of the present invention due to watching video by obtaining user's history；According to the user behaviors log data
Construct user-video rating matrix；It is completely lost decomposition, is obtained and the user-video to the user-video rating matrix
The corresponding left singular vector matrix of matrix, right singular vector matrix and diagonal matrix；According to the left singular vector matrix, institute
It states right singular vector matrix and the similar neighborhood set of the user is calculated in the diagonal matrix；According to the similar neighbour
Set and the user behaviors log data acquisition video recommendations list are occupied, and video recommendations are carried out according to the video recommendations list,
It is existing based on user in collaborative filtering-generally existing Sparse of video rating matrix so as to effectively solve
The problem of, promote the video recommendations effect of recommender system；It is nearest at the same time it can also reduce user or video during video recommendations
The time complexity that neighbour calculates, efficiently solve existing recommender system causes video recommendations result can because user volume or video explode
The problem of capable of can not timely updating, influence user experience.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, the accompanying drawings in the following description is only of the invention some
Embodiment for those of ordinary skill in the art without any creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is the schematic flow diagram for the video recommendation method that the embodiment of the present invention one provides；
Fig. 2 is the schematic flow diagram of video recommendation method provided by Embodiment 2 of the present invention；
Fig. 3 is the schematic block diagram for the video recommendation system that the embodiment of the present invention three provides；
Fig. 4 is the schematic block diagram for the video recommendation system that the embodiment of the present invention four provides；
Fig. 5 is the schematic block diagram for the terminal that the embodiment of the present invention five provides.
Specific embodiment
In being described below, for illustration and not for limitation, the tool of such as particular system structure, technology etc is proposed
Body details, to understand thoroughly the embodiment of the present invention.However, it will be clear to one skilled in the art that there is no these specific
The present invention also may be implemented in the other embodiments of details.In other situations, it omits to well-known system, device, electricity
The detailed description of road and method, in case unnecessary details interferes description of the invention.
In order to illustrate technical solutions according to the invention, the following is a description of specific embodiments.
Fig. 1 is the schematic flow diagram for the video recommendation method that the embodiment of the present invention one provides.It is shown in Figure 1, this implementation
A kind of video recommendation method that example provides may comprise steps of:
Step S101 obtains the user behaviors log data of user's history viewing video.
In the present embodiment, the user behaviors log data of the user's history viewing video include dominant feedback data and recessiveness
Feedback data.Wherein, the dominant feedback data includes but is not limited to user to the collection of video, purchase, scoring and comment
Equal behavioral datas, the stealth feedback data include but are not limited to the browsing of user, watch and during watching video
The behavioral datas such as click, search.
Step S102 constructs user-video rating matrix according to the user behaviors log data.
In the present embodiment, classification can be grouped to the user behaviors log data of user for different application scenarios, it is right
Different types of behavioral data is weighted processing, if the dominant feedback data weight of user is generally greater than explicit feedback data,
After all behaviors to user completely weight, this makes it possible to the hobbies for the viewing video for obtaining user to tend to,
Finally obtain the higher dimensional matrix that user watches video preference, i.e. user-video rating matrix.
Step S103 is completely lost decomposition to the user-video rating matrix, is obtained and the user-video matrix
Corresponding left singular vector matrix, right singular vector matrix and diagonal matrix.
In the present embodiment, step S103 is specifically included:
If the user-video rating matrix X is the matrix of a m × n, the dimensionality reduction square of the SVD singular value decomposition of X is defined
Battle array are as follows: X=U Σ VT, wherein U is the left singular vector matrix of X, and Σ is the diagonal matrix of X, VTFor right singular vector matrix, that
U, Σ, V after X decompositionTSolution mode it is as follows:
By the user-video rating matrix X transposition XTMatrix multiplication is done with X, obtains a square matrix X of n × nTX is right
Square matrix XTX carries out feature decomposition, obtains XTThe n eigenvalue λ of XiWith n feature vector vi, square matrix XTThe feature decomposition formula of X
Are as follows: (XTX)vi=λivi；Wherein, i=1,2,3.......n, n are positive integer；
By the user-video rating matrix X and X transposition XTMatrix multiplication is done, a square matrix XX of m × m is obtainedT, right
Square matrix XXTFeature decomposition is carried out, XX is obtainedTM eigenvalue λjWith m feature vector uj, square matrix XXTFeature decomposition formula
Are as follows: (XXT)uj=λjuj；Wherein, j=1,2,3.......m, m are positive integer；
According to the user-video matrix X, the left singular vector matrix U and the right singular vector matrix VTIt asks
Solution obtains each singular value σ in diagonal matrix Σk, solution formula are as follows: σk=Avi/ujOrWherein k=
1,2,3......n, n≤m.
Step S104, according to the left singular vector matrix, the right singular vector matrix and the diagonal matrix meter
Calculation obtains the similar neighborhood set of the user.
In the present embodiment, it is higher-dimension sparse matrix before user-video rating matrix dimensionality reduction, is wherein deposited in matrix and be largely
0 value, the calculated result that will lead to similarity between user or video is 0, to the nearest adjacent column of these users or video occur
Table is almost empty problem.After user-video rating matrix dimensionality reduction, user-video rating matrix is mapped to the potential of low-dimensional
In feature space, obtained user or video feature matrix are the dense matrix of low-dimensional, and the value in dense matrix is not 0 substantially,
Therefore the result based on user or video feature matrix progress similarity calculation after matrix dimensionality reduction is not substantially 0, to make
It obtains user or the corresponding arest neighbors list of video is not in for empty problem.
Step S105, according to the similar neighborhood set and the user behaviors log data acquisition video recommendations list, and root
Video recommendations are carried out according to the video recommendations list.
In the present embodiment, the similar neighborhood set includes nearest neighbor list and arest neighbors list of videos, step
S105 is specifically included:
Nearest neighbor list is obtained according to the similar neighborhood set, obtains the view of the nearest neighbor list viewing
Frequency list；And/or arest neighbors list of videos is obtained according to the similar neighborhood set；
User in the list of videos and/or the arest neighbors list of videos is rejected according to the user behaviors log data to have seen
The video seen gets the video recommendations list.
In the present embodiment, the video for having user's conception of history to see is recorded in the user behaviors log data, therefore is being obtained
It, will wherein user's history into nearest neighbor list after the list of videos and/or arest neighbors list of videos of all users viewing
After the video watched is rejected, video recommendations list can be obtained.
It preferably, in the present embodiment, can be according to the user recorded in the user behaviors log data to each type video
Comment or scoring calculate the score of each video in the video recommendations list, and according to score sequence to video recommendations
Video in list is ranked up, and user is recommended in the video recommendations list after sequence or will only be sorted forward several
Video recommendations are to user.
Above as can be seen that video recommendation method provided in this embodiment due to based on after dimensionality reduction user or video features
Matrix calculates similar with the video that user or user watched arest neighbors set, so that effective solution is existing based on cooperateing with
In filter algorithm the problem of user-video rating matrix generally existing Sparse, the video recommendations of recommender system are improved
Effect；Meanwhile the time complexity of user or the calculating of video arest neighbors during video recommendations is also reduced, it efficiently solves existing
There is recommender system that video recommendations result is caused possibly can not to timely update because user volume or video explode, influences asking for user experience
Topic.
Fig. 2 is the schematic flow diagram of video recommendation method provided by Embodiment 2 of the present invention.It is shown in Figure 2, relative to
A upper embodiment, a kind of video recommendation method provided in this embodiment construct user-according to the user behaviors log data described
Before video rating matrix further include:
Step S202 is filtered the user behaviors log data, removes the redundant data unrelated with video recommendations.
In the present embodiment, there may be program in the user behaviors log data acquisition of user's history viewing video
Log data, network status data are equal to the unrelated data of video recommendations, in the present embodiment, in building user-video rating matrix
Before, data unrelated with video recommendations in user behaviors log data are directly deleted, follow-up data treating capacity can be reduced in this way, mentioned
High follow-up data treatment effeciency.
Step S203 carries out structuring processing to filtered user behaviors log data.
In the present embodiment, the user behaviors log data are usually character string line by line, are non-structured, data knots
Structureization processing, which refers to, is cut into effective field and conversion one by one for each line character string in the user behaviors log data
It stores at the text data of specified decollator, and in the form of text or database table so that it is convenient to subsequent operation.
Step S204, to structuring, treated that user behaviors log data are normalized and missing data completion is handled.
In the present embodiment, the content of some numeric type fields is needed such as the scoring of video in structuring treated data
It limits it and normalizes value range, for example the score value by all videos is needed to be mapped between 0~1；The value of some fields is
It is empty, it is necessary to by these to be empty missing data completion, for example by the value of field be that empty value is set as 0, or by the value of field
For the average value for being set as all non-zero value.
It should be noted that step S201 and step S205~step S208 implementation in the present embodiment due to
Identical with step S101~implementation of step S105 in a upper embodiment respectively, therefore, details are not described herein.
Above as can be seen that relative to above-described embodiment, due in building user-video rating matrix in the present embodiment
Before, the user behaviors log data for first watching video to user's history are filtered, structuring is handled and normalization completion is handled etc.
Data prediction, so that subsequent builds user-video Rating Model efficiency and precision are higher, and then after improving
The matching degree of the continuous video recommended and user, improves video recommendations effect.
Fig. 3 is the schematic block diagram for the video recommendation system that the embodiment of the present invention three provides.For ease of description, only show
Part related to the present embodiment is gone out.
It is shown in Figure 3, a kind of video recommendation system 3 provided in this embodiment, comprising:
Behavioral data acquiring unit 31, for obtaining the user behaviors log data of user's history viewing video；
Rating matrix construction unit 32, for constructing user-video rating matrix according to the user behaviors log data；
Rating matrix decomposition unit 33, for being completely lost decomposition to the user-video rating matrix, obtain with it is described
The corresponding left singular vector matrix of user-video matrix, right singular vector matrix and diagonal matrix；
Similar neighborhood computing unit 34, for according to the left singular vector matrix, the right singular vector matrix and
The similar neighborhood set of the user is calculated in the diagonal matrix；
Video recommendations unit 35, for according to the similar neighborhood set and the user behaviors log data acquisition video recommendations
List, and video recommendations are carried out according to the video recommendations list.
Optionally, the rating matrix decomposition unit 33 is specifically used for:
If the user-video rating matrix X is the matrix of a m × n, the dimensionality reduction square of the SVD singular value decomposition of X is defined
Battle array are as follows: X=U Σ VT, wherein U is the left singular vector matrix of X, and Σ is the diagonal matrix of X, VTFor right singular vector matrix,
U, Σ, V after so X decompositionTSolution mode it is as follows:
By the user-video rating matrix X transposition XTMatrix multiplication is done with X, obtains a square matrix X of n × nTX is right
Square matrix XTX carries out feature decomposition, obtains XTThe n eigenvalue λ of XiWith n feature vector vi, square matrix XTThe feature decomposition formula of X
Are as follows: (XTX)vi=λivi；Wherein, i=1,2,3.......n, n are positive integer；
By the user-video rating matrix X and X transposition XTMatrix multiplication is done, a square matrix XX of m × m is obtainedT, right
Square matrix XXTFeature decomposition is carried out, XX is obtainedTM eigenvalue λjWith m feature vector uj, square matrix XXTFeature decomposition formula
Are as follows: (XXT)uj=λjuj；Wherein, j=1,2,3.......m, m are positive integer；
According to the user-video matrix X, the left singular vector matrix U and the right singular vector matrix VTIt asks
Solution obtains each singular value σ in diagonal matrix Σk, solution formula are as follows: σk=Avi/ujOrWherein k=
1,2,3......n, n≤m.
Optionally, the video recommendations unit 35 is specifically used for:
Nearest neighbor list is obtained according to the similar neighborhood set, obtains the view of the nearest neighbor list viewing
Frequency list；And/or arest neighbors list of videos is obtained according to the similar neighborhood set；
User in the list of videos and/or the arest neighbors list of videos is rejected according to the user behaviors log data to have seen
The video seen gets the video recommendations list.
Optionally, video recommendation system 3 shown in Figure 4, described further include:
Data filtering units 36 remove unrelated with video recommendations superfluous for being filtered to the user behaviors log data
Remainder evidence；
Structuring processing unit 37, for carrying out structuring processing to filtered user behaviors log data；
Completion unit 38 is normalized, for treated that user behaviors log data are normalized and missing data to structuring
Completion processing.
It should be noted that each unit in above system provided in an embodiment of the present invention, due to the method for the present invention
Embodiment is based on same design, and bring technical effect is identical as embodiment of the present invention method, and particular content can be found in this hair
Narration in bright embodiment of the method, details are not described herein again.
Therefore, it can be seen that video recommendation system provided in an embodiment of the present invention can be equally based on so that effective solution is existing
In collaborative filtering the problem of user-video rating matrix generally existing Sparse, the video for promoting recommender system is pushed away
Recommend effect；Simultaneously, additionally it is possible to reduce the time complexity of user or the calculating of video arest neighbors during video recommendations, effectively solve
Existing recommender system causes video recommendations result possibly can not timely update because user volume or video explode, and influences user experience
The problem of.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Fig. 5 is the schematic diagram for the terminal that the embodiment of the present invention five provides.As shown in figure 5, the terminal 5 of the embodiment includes:
Processor 50, memory 51 and it is stored in the computer program that can be run in the memory 51 and on the processor 50
52.The processor 50 realizes the step in above-mentioned each embodiment of the method, such as Fig. 1 institute when executing the computer program 52
The step 101 shown is to 105.Alternatively, the processor 50 realizes above-mentioned each Installation practice when executing the computer program 52
In each module/unit function, such as the function of module 31 to 35 shown in Fig. 3.
Illustratively, the computer program 52 can be divided into one or more module/units, it is one or
Multiple module/units are stored in the memory 51, and are executed by the processor 50, to complete the present invention.Described one
A or multiple module/units can be the series of computation machine program instruction section that can complete specific function, which is used for
Implementation procedure of the computer program 52 in the terminal 5 is described.For example, the computer program 52 can be divided into
Behavioral data acquiring unit 31, rating matrix construction unit 32, rating matrix decomposition unit 33, similar neighborhood computing unit 34 with
And video recommendations unit 35, each unit concrete function are as follows:
Behavioral data acquiring unit 31, for obtaining the user behaviors log data of user's history viewing video；
Rating matrix construction unit 32, for constructing user-video rating matrix according to the user behaviors log data；
Rating matrix decomposition unit 33, for being completely lost decomposition to the user-video rating matrix, obtain with it is described
The corresponding left singular vector matrix of user-video matrix, right singular vector matrix and diagonal matrix；
Similar neighborhood computing unit 34, for according to the left singular vector matrix, the right singular vector matrix and
The similar neighborhood set of the user is calculated in the diagonal matrix；
Video recommendations unit 35, for according to the similar neighborhood set and the user behaviors log data acquisition video recommendations
List, and video recommendations are carried out according to the video recommendations list.
The terminal device 5 can be the calculating such as desktop PC, notebook, palm PC and cloud server and set
It is standby.The terminal may include, but be not limited only to, processor 50, memory 51.It will be understood by those skilled in the art that Fig. 5 is only
It is the example of terminal 5, does not constitute the restriction to terminal device 5, may include than illustrating more or fewer components or group
Close certain components or different components, for example, the terminal can also include input-output equipment, it is network access equipment, total
Line etc..
Alleged processor 50 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..General processor can be microprocessor or the processor is also possible to any conventional processor
Deng.
The memory 51 can be the internal storage unit of the terminal 5, such as the hard disk or memory of terminal 5.It is described
Memory 51 is also possible to the External memory equipment of the terminal 5, such as the plug-in type hard disk being equipped in the terminal 5, intelligence
Storage card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card (Flash Card)
Deng.Further, the memory 51 can also both include the internal storage unit of the terminal 5 or set including external storage
It is standby.The memory 51 is for other programs and data needed for storing the computer program and the terminal.It is described to deposit
Reservoir 51 can be also used for temporarily storing the data that has exported or will export.
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function
Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of described device is divided into different functional unit or module, more than completing
The all or part of function of description.Each functional unit in embodiment, module can integrate in one processing unit, can also
To be that each unit physically exists alone, can also be integrated in one unit with two or more units, it is above-mentioned integrated
Unit both can take the form of hardware realization, can also realize in the form of software functional units.In addition, each function list
Member, the specific name of module are also only for convenience of distinguishing each other, the protection scope being not intended to limit this application.Above system
The specific work process of middle unit, module, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, is not described in detail or remembers in some embodiment
The part of load may refer to the associated description of other embodiments.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In embodiment provided by the present invention, it should be understood that disclosed device/terminal device and method, it can be with
It realizes by another way.For example, device described above/terminal device embodiment is only schematical, for example, institute
The division of module or unit is stated, only a kind of logical function partition, there may be another division manner in actual implementation, such as
Multiple units or components can be combined or can be integrated into another system, or some features can be ignored or not executed.Separately
A bit, shown or discussed mutual coupling or direct-coupling or communication connection can be through some interfaces, device
Or the INDIRECT COUPLING or communication connection of unit, it can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated module/unit be realized in the form of SFU software functional unit and as independent product sale or
In use, can store in a computer readable storage medium.Based on this understanding, the present invention realizes above-mentioned implementation
All or part of the process in example method, can also instruct relevant hardware to complete, the meter by computer program
Calculation machine program can be stored in a computer readable storage medium, the computer program when being executed by processor, it can be achieved that on
The step of stating each embodiment of the method.Wherein, the computer program includes computer program code, the computer program
Code can be source code form, object identification code form, executable file or certain intermediate forms etc..Computer-readable Jie
Matter may include: can carry the computer program code any entity or device, recording medium, USB flash disk, mobile hard disk,
Magnetic disk, CD, computer storage, read-only memory (ROM, Read-Only Memory), random access memory (RAM,
Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium etc..It should be noted that described
The content that computer-readable medium includes can carry out increasing appropriate according to the requirement made laws in jurisdiction with patent practice
Subtract, such as in certain jurisdictions, according to legislation and patent practice, computer-readable medium do not include be electric carrier signal and
Telecommunication signal.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features；And these are modified
Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all
It is included within protection scope of the present invention.
Claims (10)
1. a kind of video recommendation method characterized by comprising
Obtain the user behaviors log data of user's history viewing video；
User-video rating matrix is constructed according to the user behaviors log data；
Full-rank factorization is carried out to the user-video rating matrix, obtain a left side corresponding with the user-video matrix it is unusual to
Moment matrix, right singular vector matrix and diagonal matrix；
The user is calculated according to the left singular vector matrix, the right singular vector matrix and the diagonal matrix
Similar neighborhood set；
According to the similar neighborhood set and the user behaviors log data acquisition video recommendations list, and according to the video recommendations
List carries out video recommendations.
2. video recommendation method as described in claim 1, which is characterized in that it is described to the user-video rating matrix into
Row full-rank factorization obtains left singular vector matrix corresponding with the user-video matrix, right singular vector matrix and diagonal
Matrix includes:
If the user-video rating matrix X is the matrix of a m × n, the dimensionality reduction matrix of the SVD singular value decomposition of X is defined
Are as follows: X=U Σ VT, wherein U is the left singular vector matrix of X, and Σ is the diagonal matrix of X, VTFor right singular vector matrix, then
U, Σ, V after X decompositionTSolution mode it is as follows:
By the user-video rating matrix X transposition XTMatrix multiplication is done with X, obtains a square matrix X of n × nTX, to square matrix
XTX carries out feature decomposition, obtains XTThe n eigenvalue λ of XiWith n feature vector vi, square matrix XTThe feature decomposition formula of X are as follows:
(XTX)vi=λivi；Wherein, i=1,2,3.......n, n are positive integer；
By the user-video rating matrix X and X transposition XTMatrix multiplication is done, a square matrix XX of m × m is obtainedT, to square matrix
XXTFeature decomposition is carried out, XX is obtainedTM eigenvalue λjWith m feature vector uj, square matrix XXTFeature decomposition formula are as follows:
(XXT)uj=λjuj；Wherein, j=1,2,3.......m, m are positive integer；
According to the user-video matrix X, the left singular vector matrix U and the right singular vector matrix VTSolution obtains
Each singular value σ in diagonal matrix Σk, solution formula are as follows: σk=Avi/ujOrWherein k=1,2,
3......n, n≤m.
3. video recommendation method as described in claim 1, which is characterized in that described constructed according to the user behaviors log data is used
Before family-video rating matrix further include:
The user behaviors log data are filtered, the redundant data unrelated with video recommendations is removed；
Structuring processing is carried out to filtered user behaviors log data；
To structuring, treated that user behaviors log data are normalized and missing data completion processing.
4. video recommendation method as described in claim 1, which is characterized in that described according to the similar neighborhood set and described
User behaviors log data acquisition video recommendations list includes:
Nearest neighbor list is obtained according to the similar neighborhood set, obtains the video column of the nearest neighbor list viewing
Table；And/or arest neighbors list of videos is obtained according to the similar neighborhood set；
User in the list of videos and/or the arest neighbors list of videos is rejected according to the user behaviors log data to have already viewed
Video, get the video recommendations list.
5. a kind of video recommendation system characterized by comprising
Behavioral data acquiring unit, for obtaining the user behaviors log data of user's history viewing video；
Rating matrix construction unit, for constructing user-video rating matrix according to the user behaviors log data；
Rating matrix decomposition unit obtains and the user-for carrying out full-rank factorization to the user-video rating matrix
The corresponding left singular vector matrix of video matrix, right singular vector matrix and diagonal matrix；
Similar neighborhood computing unit, for according to the left singular vector matrix, the right singular vector matrix and described right
The similar neighborhood set of the user is calculated in angular moment battle array；
Video recommendations unit is used for according to the similar neighborhood set and the user behaviors log data acquisition video recommendations list,
And video recommendations are carried out according to the video recommendations list.
6. video recommendation system as claimed in claim 5, which is characterized in that the rating matrix decomposition unit is specifically used for:
If the user-video rating matrix X is the matrix of a m × n, the dimensionality reduction matrix of the SVD singular value decomposition of X is defined
Are as follows: X=U Σ VT, wherein U is the left singular vector matrix of X, and Σ is the diagonal matrix of X, VTFor right singular vector matrix, then
U, Σ, V after X decompositionTSolution mode it is as follows:
By the user-video rating matrix X transposition XTMatrix multiplication is done with X, obtains a square matrix X of n × nTX, to square matrix
XTX carries out feature decomposition, obtains XTThe n eigenvalue λ of XiWith n feature vector vi, square matrix XTThe feature decomposition formula of X are as follows:
(XTX)vi=λivi；Wherein, i=1,2,3.......n, n are positive integer；
By the user-video rating matrix X and X transposition XTMatrix multiplication is done, a square matrix XX of m × m is obtainedT, to square matrix
XXTFeature decomposition is carried out, XX is obtainedTM eigenvalue λjWith m feature vector uj, square matrix XXTFeature decomposition formula are as follows:
(XXT)uj=λjuj；Wherein, j=1,2,3.......m, m are positive integer；
According to the user-video matrix X, the left singular vector matrix U and the right singular vector matrix VTSolution obtains
Each singular value σ in diagonal matrix Σk, solution formula are as follows: σk=Avi/ujOrWherein k=1,2,
3......n, n≤m.
7. video recommendation system as claimed in claim 5, which is characterized in that further include:
Data filtering units remove the redundant data unrelated with video recommendations for being filtered to the user behaviors log data；
Structuring processing unit, for carrying out structuring processing to filtered user behaviors log data；
Normalize completion unit, for structuring treated user behaviors log data are normalized and missing data completion at
Reason.
8. video recommendation system as claimed in claim 5, which is characterized in that the video recommendations unit is specifically used for:
Nearest neighbor list is obtained according to the similar neighborhood set, obtains the video column of the nearest neighbor list viewing
Table；And/or arest neighbors list of videos is obtained according to the similar neighborhood set；
User in the list of videos and/or the arest neighbors list of videos is rejected according to the user behaviors log data to have already viewed
Video, get the video recommendations list.
9. a kind of terminal, including memory, processor and storage can be run in the memory and on the processor
Computer program, which is characterized in that the processor is realized when executing the computer program as Claims 1-4 is any
The step of item the method.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In when the computer program is executed by processor the step of any one of such as Claims 1-4 of realization the method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710850056.XA CN109947983A (en) | 2017-09-19 | 2017-09-19 | Video recommendation method, system, terminal and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710850056.XA CN109947983A (en) | 2017-09-19 | 2017-09-19 | Video recommendation method, system, terminal and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109947983A true CN109947983A (en) | 2019-06-28 |
Family
ID=67003978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710850056.XA Pending CN109947983A (en) | 2017-09-19 | 2017-09-19 | Video recommendation method, system, terminal and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109947983A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866181A (en) * | 2019-10-12 | 2020-03-06 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, device and storage medium |
| CN112135193A (en) * | 2020-09-24 | 2020-12-25 | 湖南快乐阳光互动娱乐传媒有限公司 | Video recommendation method and device |
| CN112380389A (en) * | 2020-11-18 | 2021-02-19 | 中信银行股份有限公司 | Video recommendation method and device, electronic equipment and readable storage medium |
| CN116028727A (en) * | 2023-03-30 | 2023-04-28 | 南京邮电大学 | Video recommendation method based on image data processing |
| CN117235366A (en) * | 2023-09-19 | 2023-12-15 | 北京学说科技有限公司 | Collaborative recommendation method and system based on content relevance |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040158497A1 (en) * | 2003-02-06 | 2004-08-12 | Brand Matthew E. | On-line recommender system |
| CN103620592A (en) * | 2011-04-19 | 2014-03-05 | 诺基亚公司 | Method and apparatus for flexible diversification of recommendation results |
| CN105426392A (en) * | 2015-10-28 | 2016-03-23 | 浪潮软件集团有限公司 | Collaborative filtering recommendation method and system |
| US20160147768A1 (en) * | 2014-11-25 | 2016-05-26 | Samsung Electronics Co., Ltd. | Device and method for providing media resource |
| CN106649540A (en) * | 2016-10-26 | 2017-05-10 | Tcl集团股份有限公司 | Video recommendation method and system |
-
2017
- 2017-09-19 CN CN201710850056.XA patent/CN109947983A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040158497A1 (en) * | 2003-02-06 | 2004-08-12 | Brand Matthew E. | On-line recommender system |
| CN103620592A (en) * | 2011-04-19 | 2014-03-05 | 诺基亚公司 | Method and apparatus for flexible diversification of recommendation results |
| US20160147768A1 (en) * | 2014-11-25 | 2016-05-26 | Samsung Electronics Co., Ltd. | Device and method for providing media resource |
| CN105426392A (en) * | 2015-10-28 | 2016-03-23 | 浪潮软件集团有限公司 | Collaborative filtering recommendation method and system |
| CN106649540A (en) * | 2016-10-26 | 2017-05-10 | Tcl集团股份有限公司 | Video recommendation method and system |
Non-Patent Citations (1)
| Title |
|---|
| 吴奕: "一种改进的协同过滤技术在视频推荐中的应用", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866181A (en) * | 2019-10-12 | 2020-03-06 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, device and storage medium |
| CN110866181B (en) * | 2019-10-12 | 2022-04-22 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, device and storage medium |
| CN112135193A (en) * | 2020-09-24 | 2020-12-25 | 湖南快乐阳光互动娱乐传媒有限公司 | Video recommendation method and device |
| CN112380389A (en) * | 2020-11-18 | 2021-02-19 | 中信银行股份有限公司 | Video recommendation method and device, electronic equipment and readable storage medium |
| CN116028727A (en) * | 2023-03-30 | 2023-04-28 | 南京邮电大学 | Video recommendation method based on image data processing |
| CN116028727B (en) * | 2023-03-30 | 2023-08-18 | 南京邮电大学 | Video recommendation method based on image data processing |
| CN117235366A (en) * | 2023-09-19 | 2023-12-15 | 北京学说科技有限公司 | Collaborative recommendation method and system based on content relevance |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109947983A (en) | Video recommendation method, system, terminal and computer readable storage medium | |
| US20210173711A1 (en) | Integrated value chain risk-based profiling and optimization | |
| CN111506801B (en) | Sequencing method and device for application App neutron application | |
| CN110020122A (en) | A kind of video recommendation method, system and computer readable storage medium | |
| CN110135942A (en) | Products Show method, apparatus and computer readable storage medium | |
| CN107786943A (en) | A kind of tenant group method and computing device | |
| CN109471978B (en) | Electronic resource recommendation method and device | |
| CN113918832B (en) | Graph convolution collaborative filtering recommendation system based on social relationship | |
| CN107704070A (en) | Using method for cleaning, device, storage medium and electronic equipment | |
| CN116010684A (en) | Article recommendation method, device and storage medium | |
| CN113918834B (en) | Graph convolution collaborative filtering recommendation method fusing social relations | |
| CN109087138A (en) | Data processing method and system, computer system and readable storage medium storing program for executing | |
| WO2023087914A1 (en) | Method and apparatus for selecting recommended content, and device, storage medium and program product | |
| CN110473073A (en) | The method and device that linear weighted function is recommended | |
| CN109977264A (en) | A kind of user's recommended method, system, server and storage medium based on first similarity of paths | |
| Truong et al. | Exploring cross-modality utilization in recommender systems | |
| Park et al. | On the power of gradual network alignment using dual-perception similarities | |
| CN112507098B (en) | Question processing method, question processing device, electronic equipment, storage medium and program product | |
| Provalov et al. | Synevarec: A framework for evaluating recommender systems on synthetic data classes | |
| CN113761350A (en) | Data recommendation method, related device and data recommendation system | |
| CN112989170A (en) | Keyword matching method applied to information search, information search method and device | |
| CN110276283B (en) | Picture identification method, target identification model training method and device | |
| Gupta et al. | Framework to Analyze Customer’s Feedback in Smartphone Industry Using Opinion Mining. | |
| CN109886299B (en) | User portrait method and device, readable storage medium and terminal equipment | |
| CN109872183A (en) | Intelligent Service evaluation method, computer readable storage medium and terminal device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190628 |