CN109063052A - A kind of personalized recommendation method and device based on time entropy - Google Patents
A kind of personalized recommendation method and device based on time entropy Download PDFInfo
- Publication number
- CN109063052A CN109063052A CN201810794746.2A CN201810794746A CN109063052A CN 109063052 A CN109063052 A CN 109063052A CN 201810794746 A CN201810794746 A CN 201810794746A CN 109063052 A CN109063052 A CN 109063052A
- Authority
- CN
- China
- Prior art keywords
- user
- project
- active user
- time entropy
- past
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 description 18
- 238000005516 engineering process Methods 0.000 description 7
- 238000001914 filtration Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000036651 mood Effects 0.000 description 2
- 101150096839 Fcmr gene Proteins 0.000 description 1
- 101000948324 Homo sapiens Protein CutA Proteins 0.000 description 1
- 102100036046 Protein CutA Human genes 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000005612 types of electricity Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present invention provides a kind of personalized recommendation method and device based on time entropy, which comprises clusters to existing user and project score information, obtains the user for including in the class cluster center and each class cluster of cluster and project information；If it is determined that active user is not belonging to that system user newly is added, then the active user is calculated to item object time entropy to the different scoring records of project according to the association attributes characteristic information of the active user and user；According to the calculated active user to item object time entropy, the user type of the active user is determined；Pattern match is carried out at the class cluster center of cluster according to the user type of the active user, to carry out the recommendation of diversity project information for the active user.The embodiment of the present invention can be improved information recommendation accuracy and recommend diversity.
Description
Technical field
The present invention relates to internet intelligent technical field of information recommendation more particularly to a kind of personalization based on time entropy to push away
Recommend method and device.
Background technique
With the rise of Internet technology and the fast development of information technology, internet produces a large amount of data information.
The chart-information for the portion " internet generates data in one minute " issued by Excelcom company, Facebook common property known to us
Raw 701,389 accounts log in, Netflix share 69,444 small durations video watched, Snapchat has shared 527,760
Upper 51,000 app of photo, App Store are downloaded, Linkedin creates more than 120 new account, Twitter and issues
347,222 newly push away text, Instagram has issued 28,194 new photos, Google produces the request of 2,400,000 new search,
So that internet is from the era trend of original absence of information information overload (Information overload), this but also with
Want quickly and accurately to find its interested information from massive information library and become more difficult in family.
In face of problem of information overload, ordinary user can not often be followed.Scientist is in order to preferably meet the information of user
Demand proposes recommender system technology, and the technology is by by machine learning, data mining, user behavior and human-computer interaction etc.
The technology of multiple fields is combined, and uses large-scale parallel data processing shelf, and then is quickly and accurately every use
Family provides individual info service.Collaborative filtering is a kind of effective technology that can generate personalized recommendation, is in various recommendations
It is all used widely in system, basic task is to match user according to similar preference, may be liked with recommended user
Project.Collaborative filtering can be generally divided into based on memory and based on model.Wherein, collaborative filtering memory-based can divide again
For based on user and based on project.The former is to calculate similarity between user, is obtained similar with target user's interest preference nearest
Neighbour carries out prediction recommendation based on this.
However, traditional Collaborative Filtering Recommendation Algorithm is when for user's recommended project, usually only select based on user or
Project-based recommended method is that target user recommends, and this way of recommendation has only selected user to believe the scoring of project
Breath, and the interest for having ignored user can be become with the variation of time by other influences such as mood, friend and fashions
Change, that is to say, that the interest of user may only focus on one or several projects over a period to come, i.e. interest migrates.Therefore,
The recommendation of single type is not able to satisfy the practicability demand of other diversity user.
Summary of the invention
The embodiment of the present invention provides a kind of personalized recommendation method and device based on time entropy, to improve information recommendation standard
Exactness and recommendation diversity.
On the one hand, the embodiment of the invention provides a kind of personalized recommendation methods based on time entropy, which comprises
Existing user and project score information are clustered, obtains in the class cluster center and each class cluster of cluster and includes
User and project information；
If it is determined that active user is not belonging to that system user newly is added, then believed according to the association attributes feature of the active user
Breath and user calculate the active user to item object time entropy to the different scoring records of project；
According to the calculated active user to item object time entropy, the user type of the active user is determined；
Pattern match is carried out at the class cluster center of cluster according to the user type of the active user, to be described
Active user carries out the recommendation of diversity project information.
On the other hand, the embodiment of the invention provides a kind of personalized recommendation device based on time entropy, described device packets
It includes:
Cluster cell, for being clustered to existing user and project score information, obtain cluster class cluster center and
The user for including in each class cluster and project information；
Computing unit, for if it is determined that active user is not belonging to that system user newly is added, then according to the active user's
Association attributes characteristic information and user calculate the active user to item object time entropy to the different scoring records of project；
Judging unit, for, to item object time entropy, determining the active user according to the calculated active user
User type；
Matching unit, for carrying out mode at the class cluster center of cluster according to the user type of the active user
Match, to carry out the recommendation of diversity project information for the active user.
Above-mentioned technical proposal, which has the following beneficial effects:, carries out user using the personalized recommendation method based on time entropy
Recommend, on the one hand, user and project information are clustered and computational item purpose popularity, efficiency, the standard of recommendation can be improved
Exactness and the cold start-up for solving the problems, such as user.On the other hand, use can be efficiently used to item object time entropy by calculating user
More interest at family, to improve the diversity of recommended method.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of personalized recommendation method flow chart based on time entropy of the embodiment of the present invention；
Fig. 2 is a kind of personalized recommendation apparatus structure schematic diagram based on time entropy of the embodiment of the present invention；
Fig. 3 is personalized recommendation apparatus structure schematic diagram of the another kind of the embodiment of the present invention based on time entropy；
Fig. 4 is the overall flow figure of personalized recommendation method of the application example of the present invention based on time entropy；
Fig. 5 is that personalized recommendation method of the application example of the present invention based on time entropy and other proposed algorithms recommend efficiency pair
Than figure；
Fig. 6 is personalized recommendation method diversity comparison diagram of the application example of the present invention based on time entropy.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, being a kind of personalized recommendation method flow chart based on time entropy of the embodiment of the present invention, the method
Include:
101, existing user and project score information are clustered, is obtained in the class cluster center and each class cluster of cluster
The user for including and project information；
102, if it is determined that active user is not belonging to that system user newly is added, then special according to the association attributes of the active user
Reference breath and user calculate the active user to item object time entropy to the different scoring records of project；
103, the user class of the active user is determined to item object time entropy according to the calculated active user
Type；
104, pattern match is carried out at the class cluster center of cluster according to the user type of the active user, to be
The active user carries out the recommendation of diversity project information.
Preferably, the method also includes:
If it is determined that active user is new addition system user, it is using following project popularity calculation that ranking is preceding
N number of project information recommends the active user:
Wherein, ipopFor project popularity, UiUser's set that project i was evaluated in recommender system is represented, I is to recommend system
All items number in system.
Preferably, according to user interest preference, determine that the user type of the active user is one of following four seed type:
Like recently and also likes in the past, likes recently but do not like in the past, do not like recently but like in the past, do not like recently and mistake
It goes also not liking.
Preferably, the user type according to the active user carries out mode at the class cluster center of cluster
Match, to carry out the recommendation of diversity project information for the active user, comprising:
The following two kinds interest preference mode according to the user type of the active user at the class cluster center of cluster
Middle carry out pattern match: regular interest preference mode and non-regularity interest preference mode；Wherein, the regular interest is inclined
Good model is to like recently and also like in the past；The non-regularity interest preference mode includes: to like recently but do not like in the past
Vigorously, it does not like but likes in the past recently, do not like recently and do not like in the past yet.
Preferably, described that the different scorings of project are remembered according to the association attributes characteristic information and user of the active user
Record calculates the active user to item object time entropy using following time entropy formula:
Wherein, H is time entropy, and n is time interval number, scoreiScoring for user to project i, miRepresent a time
Interior all users are spaced to the scoring summation of project i, M refers to all miSummation.
Corresponding to above method embodiment, as shown in Fig. 2, being pushed away for a kind of personalization based on time entropy of the embodiment of the present invention
Apparatus structure schematic diagram is recommended, described device includes:
Cluster cell 21 obtains the class cluster center of cluster for clustering to existing user and project score information
With the user and project information for including in each class cluster；
Computing unit 22, for if it is determined that active user is not belonging to that system user newly is added, then according to the active user
Association attributes characteristic information and user the active users are calculated to item object time entropy to the different scoring records of project；
Judging unit 23, for, to item object time entropy, determining the current use according to the calculated active user
The user type at family；
Matching unit 24, for carrying out mode at the class cluster center of cluster according to the user type of the active user
Matching, to carry out the recommendation of diversity project information for the active user.
Preferably, as shown in figure 3, being shown for another kind of the embodiment of the present invention based on the personalized recommendation apparatus structure of time entropy
It is intended to, described device not only includes: cluster cell 21, computing unit 22, judging unit 23, matching unit 24, and described device is also
Include:
Recommendation unit 25 is used for if it is determined that active user is new addition system user, using following project popularity computation
The preceding N number of project information of ranking is recommended the active user by mode:
Wherein, ipopFor project popularity, UiUser's set that project i was evaluated in recommender system is represented, I is to recommend system
All items number in system.
Preferably, the judging unit 23 is specifically used for determining the user of the active user according to user interest preference
Type is one of following four seed type: like recently and also like in the past, like recently but do not like in the past, do not like recently but
Like in the past, do not like recently and does not also like in the past.
Preferably, the matching unit 24, specifically for according to the user type of the active user in cluster described in
Pattern match is carried out in the following two kinds interest preference mode at class cluster center: regular interest preference mode and non-regularity interest
Preference pattern；Wherein, the regular interest preference mode is to like recently and also like in the past；The non-regularity interest is inclined
Good model includes: to like recently but do not like in the past, do not like recently but like in the past, do not like recently and also do not like in the past
Vigorously.
Preferably, the computing unit 22, specifically for the association attributes characteristic information and use according to the active user
Family calculates the active user to item object time entropy using following time entropy formula to the different scoring records of project:
Wherein, H is time entropy, and n is time interval number, scoreiScoring for user to project i, miRepresent a time
Interior all users are spaced to the scoring summation of project i, M refers to all miSummation.
Above-mentioned technical proposal of the embodiment of the present invention, which has the following beneficial effects:, utilizes the personalized recommendation side based on time entropy
Method recommends user, on the one hand, clusters and computational item purpose popularity to user and project information, can be improved and push away
Efficiency, accuracy and the cold start-up for solving the problems, such as user recommended.It on the other hand, can be with to item object time entropy by calculating user
The more interest for efficiently using user, to improve the diversity of recommended method.
Above embodiments of the present invention propose a kind of personalized recommendation method (Personalized based on time entropy
Time Collaborative Filtering, PTCF) it is that user recommends.This method is first to existing in recommender system
User and project score information cluster, to obtain corresponding cluster centre and different class cluster information；Then determine mesh
Whether mark user belongs to new addition system user, if not then according to the association attributes characteristic information of user and user to project
Different scoring records calculate users to item object time entropy, and determine that user belongs in four kinds of user types which kind of use
Family, so in recommender system target user carry out pattern match, thus use item diversity calculation formula for user into
Row diversity is recommended；Otherwise use project popularity calculation by the preceding N number of project recommendation of ranking to target user.
The main recommendation mechanisms of the above-mentioned this personalized recommendation method PTCF based on time entropy are: in proposed algorithm
In selection, the collaborative filtering based on cluster is not only selected to be recommended, while in view of the interest of user can be with
The variation of various factors and change so as to user carry out diversity recommendation.This way of recommendation not only compensates for recommendation system
It unites in the original recommendation efficiency for calculating whole score informations and being recommended, while the diversification scored using user project
Information, to effectively improve the recommendation accuracy of the recommended method and recommend diversity.
The research and application of the to the effect that personalized recommendation method PTCF based on time entropy of the embodiment of the present invention, it is main
To include original user and project score information be clustered, how to calculate time entropy that target user scores to disparity items,
How whether target user carry out classification and pattern match and be that new addition system user progress is corresponding for target user
Recommend.Its technical solution used are as follows: 1) by gathering with RLPSO_KM clustering algorithm to user and project score information
Class, thus obtain cluster class cluster center and corresponding class cluster information；2) by introducing the definition of time entropy, added to calculate
Enter the target user of system to different item object time entropy, and then classify and pattern match for target user, thus for it
Carry out validity recommendation；3) by introducing calculating project popularity mode, to be pushed away for the new target user that system is added
It recommends.
Technical solution used in the embodiment of the present invention is a kind of personalized recommendation method based on time entropy, the reality of this method
It is existing that steps are as follows:
(1) user and project score information are clustered.Firstly, score information of the user to project is handled；Then
Using Kmeans clustering algorithm RLPSO_KM is improved, to treated, information is clustered, including initial cluster center and just
The setting of the series of parameters such as the number of iterations of beginning clusters number and cluster；Finally export cluster class cluster center and each class cluster
The user for inside including and project information.
(2) personalized recommendation based on time entropy.The interest preference of different user is often different, and some users like always
The film of same type, and for other users then just on the contrary, they may be with theirs to the love style of film
Mood, friend and faddish trend and change.
Here according to user to the score information of project, the interest preference of user is divided by we likes (Likes) and not
Like (Dislikes).Similarly if measured with the time, it is classified as nearest (Recent) and passes by the time of (Past),
So the interest preference of user is simply divided into four seed types by us.First seed type is RecentLikes；
The interest preference of PastLikes, this kind of user have regularity, that is to say, that such user usually only likes a type of electricity
Shadow, and the duration for liking a kind of film types is longer, and general a very long time will not change.Second of user
Type is RecentLikes；PastDislikes, third seed type are RecentDislikes；PastLikes.For these
For user, their interest preference changes with the change of time.Finally one is RecentDislikes；
PastDislikes, diversified trend is presented in the project information that these user's evaluations are crossed, while randomness and nothing is presented in project
Regularity.
By classifying to user, the type of user is divided into both of which by us, respectively the first mode (first
Class user), second of mode (second and third, four class users).The value and temporal information implied due to project means that user belongs to
In the user of which kind of mode.For example, if being frequently visited by the user the film of same type, it is believed that he is to follow the first mould
The user of formula.This mode is very simple, and equally distributed trend is presented in same type of film.On the contrary, if the user belongs to
Second of mode can then be followed without regularity.It is inspired by this, we have proposed a kind of new ant algorithms excavation sides based on time entropy
Method, this method can be used to measure the Annual distribution of user's history scoring record, and the calculation formula for defining time entropy is as follows:
In formula (1), n is time interval number, and user is score to the scoring of project ii, miIt represents in a time interval
All users refer to all m to the scoring summation of project i, MiSummation.
A usual project contains multiple labels, project can be briefly described by label.For a film, this
A little labels can represent the type of film.For a user, he can comment on identical label, i.e., one in different times
Label can be subordinated to the period of scoring several times.The confusion degree that time entropy can be stabbed with measure time, the value the high, shows user
More like this label.
It as shown in table 1, include score information of 4 users to 8 projects in the table.ui(i=1,2,3,4) is user's letter
Breath set, ik(k=1,2 ..., 8) represents project information set.User between 1 to 5, goes back the score value of project in table
User is set out to the access times of project and to the score information of the affiliated label of project.
1 user of table-project score information
Table 1 User-item rating
As shown in table 2, there are 5 time intervals in table.For user u1, include 3 Annual distribution records.For user u2,
It also includes 3 Annual distribution records.
2 user of table is to tag1Score information
Table 2 Users rating for tag1
But for user u1With user u2, prefer which user tag1It is their time entropy below:
Due toSo user u1Than user u2Prefer tag1.This and user u1Than user u2To tag1Scoring
The higher fact is consistent.
Diversity is recommended recommendatory to improve by using the similitude and difference of coordinating the project in bulleted list
Recommendation efficiency attracted a large amount of concern.A kind of multifarious method of general criterion is maximization different item as far as possible
Purpose summation.Therefore, the project given for one, the diversity formula that we set it are defined as follows:
In formula (2), fr (i) index label tag1The frequency occurred in set, z (i) are label tag1Institute in candidate collection
There is the sum of the similarity of label, α is adjustment factor, and it is 2 that we, which choose the value, here.
In formula (4), item ruiAnd rujScoring of the user u to project i and project j is respectively represented,Item was evaluated to be all
The user of mesh j to the average score of project j,Average score for all users for evaluating project i to project i, UijIt is same
When to project i and project j scoring user gather, the value of sim (i, j) is between [- 1,1].
(3) project popularity is calculated.For the target user being newly added in system, we then recommend most popular with users
N number of project is to target user.The popularity formula that we set project in recommender system is defined as follows:
In formula (5), UiUser's set that project i was evaluated in recommender system is represented, I is all items in recommender system
Mesh.
To sum up, by clustering to score information of the user in collaborative filtering to project, then pass through calculating
User calculates the affiliated mode of user to item object time entropy and then is that user carries out diversity recommendation, and in system not to be added
The user of system carries out the recommendation of top n project.In this way, on the one hand by using the neighbour for calculating user in class cluster
Domain improves the recommendation efficiency of recommender system；On the other hand, by calculating user to item object time entropy, so that the interest of user
It is greatly embodied, compensates for the unicity of conventional recommendation systems, so that recommender system has better diversity.
For the purpose of the present invention, technical solution and feature is more clearly understood, below in conjunction with specific embodiment, and join
According to attached drawing, further refinement explanation is carried out to the present invention.The overall flow figure of personalized recommendation method based on time entropy is such as
Shown in Fig. 4.
Each step is described as follows:
(1) propose that a kind of RLPSO_KM clustering algorithm that is based on clusters user and score information, so that recommender system
Recommendation efficiency be improved.
(2) it proposes a kind of personalized recommendation method based on time entropy, introduces time entropy calculation, so that with
The diversified interest at family has obtained very big embodiment.
(3) mode for proposing a kind of calculating project popularity is that the new system user that is added is recommended, and avoids recommendation system
The cold start-up problem of project in system.
Experimental situation is as follows:
Application example of the present invention passes through the reality of the experimental verification personalized recommendation method proposed in this paper based on time entropy
Effect, experimental situation be win7 (64) host, 8G memory, 1T hard disk, the data set used for MovieLens (10M) and from
The data of Douban.com crawl, and the average value that algorithm 10 times are tested is verified into the standard of recommendation as final experimental result
True property, while comparing the diversity of personalized recommendation method and other diversity recommended methods based on time entropy.
Firstly, application example of the present invention clusters the score information of user and project.For score information carry out to
Quantization or characterization, so that score information meets the format of cluster input；Then by using improved clustering algorithm
To treated, information clusters RLPSO_KM, selection and cluster including initial cluster center and initial clustering number
The setting of the series of parameters such as the number of iterations；It finally exports and is wrapped in each class cluster center and each class cluster of clustered algorithm cluster
The user contained and project information.
Then, the personalized recommendation method based on time entropy is established.By introducing the calculation of time entropy, user is calculated
To item object time entropy, and then classify to user and pattern match, is finally that user carries out diversity recommendation.The recommendation is calculated
The recommendation efficiency comparative of method and other algorithms figure is as shown in Figure 5 and Figure 6.
Finally, calculating project popularity.For the new user that system is not added, by computational item purpose popularity, and select
Take wherein the highest N number of project of popularity be user recommend, to solve the problems, such as the cold start-up of user in system.
Method proposed by the present invention is compared with other algorithms from Fig. 5, can clearly be seen proposed by the present invention
No matter the experimental data set that PTCF is selected is MovieLens or Douban data set, several relative to PMF, BPMF, SVD, McoC
For kind of algorithm, all more take advantage on accurate rate, accuracy rate on MovieLens and Douban data set relative to
1.04% and 1.07% has been respectively increased in N=10 in MCoC algorithm.Meanwhile no matter the Top-N value that we choose is 10,20 Hes
Any one in 50, the present invention propose that the MAP value of method is all apparently higher than other several algorithms, and this also illustrates the present invention to mention
Method PTCF is recommending the advantage in efficiency out.
We are proposed by the present invention it is found that by comparing in the case where different recommended project list numbers from Fig. 6
PTCF method no matter experimental data set select or when recommendation items mesh number be 10,20 and 50 when, diversity is protected substantially
It holds unanimously, does not substantially change.Simultaneously for CUTA algorithm, as N=10, personalized recommendation proposed by the present invention
The diversity of method improves 4 times under MovieLens data set, improves 1.05% for CUTATime algorithm.
In summary it tests, application example of the present invention pushes away user using the personalized recommendation method based on time entropy
It recommends, on the one hand, user and project information are clustered and computational item purpose popularity, the efficiency, accurate of recommendation can be improved
Spend and solve the problems, such as the cold start-up of user.On the other hand, user can be efficiently used to item object time entropy by calculating user
More interest, to improve the diversity of recommended method.
It should be understood that the particular order or level of the step of during disclosed are the examples of illustrative methods.Based on setting
Count preference, it should be appreciated that in the process the step of particular order or level can be in the feelings for the protection scope for not departing from the disclosure
It is rearranged under condition.Appended claim to a method is not illustratively sequentially to give the element of various steps, and not
It is to be limited to the particular order or level.
In above-mentioned detailed description, various features are combined together in single embodiment, to simplify the disclosure.No
This published method should be construed to reflect such intention, that is, the embodiment of theme claimed needs to compare
The more features of the feature clearly stated in each claim.On the contrary, as appended claims is reflected
Like that, the present invention is in the state fewer than whole features of disclosed single embodiment.Therefore, appended claims
It is hereby expressly incorporated into detailed description, wherein each claim is used as alone the individual preferred embodiment of the present invention.
For can be realized any technical staff in the art or using the present invention, above to disclosed embodiment into
Description is gone.To those skilled in the art；The various modifications mode of these embodiments will be apparent from, and this
The General Principle of text definition can also be suitable for other embodiments on the basis of not departing from the spirit and scope of the disclosure.
Therefore, the disclosure is not limited to embodiments set forth herein, but most wide with principle disclosed in the present application and novel features
Range is consistent.
Description above includes the citing of one or more embodiments.Certainly, in order to describe above-described embodiment and description portion
The all possible combination of part or method is impossible, but it will be appreciated by one of ordinary skill in the art that each implementation
Example can do further combinations and permutations.Therefore, embodiment described herein is intended to cover fall into the appended claims
Protection scope in all such changes, modifications and variations.In addition, with regard to term used in specification or claims
The mode that covers of "comprising", the word is similar to term " includes ", just as " including " solved in the claims as transitional word
As releasing.In addition, the use of any one of specification in claims term "or" being to indicate " non-exclusionism
Or ".
Those skilled in the art will also be appreciated that the various illustrative components, blocks that the embodiment of the present invention is listed
(illustrative logical block), unit and step can by electronic hardware, computer software, or both knot
Conjunction is realized.For the replaceability (interchangeability) for clearly showing that hardware and software, above-mentioned various explanations
Property component (illustrative components), unit and step universally describe their function.Such function
It can be that the design requirement for depending on specific application and whole system is realized by hardware or software.Those skilled in the art
Can be can be used by various methods and realize the function, but this realization is understood not to for every kind of specific application
Range beyond protection of the embodiment of the present invention.
Various illustrative logical blocks or unit described in the embodiment of the present invention can by general processor,
Digital signal processor, specific integrated circuit (ASIC), field programmable gate array or other programmable logic devices, discrete gate
Or transistor logic, discrete hardware components or above-mentioned any combination of design carry out implementation or operation described function.General place
Managing device can be microprocessor, and optionally, which may be any traditional processor, controller, microcontroller
Device or state machine.Processor can also be realized by the combination of computing device, such as digital signal processor and microprocessor,
Multi-microprocessor, one or more microprocessors combine a digital signal processor core or any other like configuration
To realize.
The step of method described in the embodiment of the present invention or algorithm can be directly embedded into hardware, processor execute it is soft
The combination of part module or the two.Software module can store in RAM memory, flash memory, ROM memory, EPROM storage
Other any form of storaging mediums in device, eeprom memory, register, hard disk, moveable magnetic disc, CD-ROM or this field
In.Illustratively, storaging medium can be connect with processor, so that processor can read information from storaging medium, and
It can be to storaging medium stored and written information.Optionally, storaging medium can also be integrated into the processor.Processor and storaging medium can
To be set in asic, ASIC be can be set in user terminal.Optionally, processor and storaging medium also can be set in
In different components in the terminal of family.
In one or more exemplary designs, above-mentioned function described in the embodiment of the present invention can be in hardware, soft
Part, firmware or any combination of this three are realized.If realized in software, these functions be can store and computer-readable
On medium, or it is transferred on a computer readable medium in the form of one or more instructions or code forms.Computer readable medium includes electricity
Brain storaging medium and convenient for so that computer program is allowed to be transferred to from a place telecommunication media in other places.Storaging medium can be with
It is that any general or special computer can be with the useable medium of access.For example, such computer readable media may include but
It is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disc storages, disk storage or other magnetic storage devices or other
What can be used for carry or store with instruct or data structure and it is other can be by general or special computer or general or specially treated
The medium of the program code of device reading form.In addition, any connection can be properly termed computer readable medium, example
Such as, if software is to pass through a coaxial cable, fiber optic cables, double from a web-site, server or other remote resources
Twisted wire, Digital Subscriber Line (DSL) are defined with being also contained in for the wireless way for transmitting such as example infrared, wireless and microwave
In computer readable medium.The disk (disk) and disk (disc) includes compress disk, radium-shine disk, CD, DVD, floppy disk
And Blu-ray Disc, disk is usually with magnetic replicate data, and disk usually carries out optically replicated data with laser.Combinations of the above
Also it may be embodied in computer readable medium.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (10)
1. a kind of personalized recommendation method based on time entropy, which is characterized in that the described method includes:
Existing user and project score information are clustered, the use for including in the class cluster center and each class cluster of cluster is obtained
Family and project information；
If it is determined that active user is not belonging to that system user newly is added, then according to the association attributes characteristic information of the active user and
User calculates the active user to item object time entropy to the different scoring records of project；
According to the calculated active user to item object time entropy, the user type of the active user is determined；
Pattern match is carried out at the class cluster center of cluster according to the user type of the active user, thus to be described current
User carries out the recommendation of diversity project information.
2. the personalized recommendation method as described in claim 1 based on time entropy, which is characterized in that the method also includes:
If it is determined that active user is new addition system user, it is using following project popularity calculation that ranking is preceding N number of
Project information recommends the active user:
Wherein, ipopFor project popularity, UiUser's set that project i was evaluated in recommender system is represented, I is in recommender system
All items number.
3. the personalized recommendation method as described in claim 1 based on time entropy, which is characterized in that according to user interest preference,
The user type for determining the active user is one of following four seed type: like recently and also like in the past, like recently but
It does not like in the past, like but like in the past recently, do not like recently and do not like in the past yet.
4. the personalized recommendation method as claimed in claim 3 based on time entropy, which is characterized in that described according to the current use
The user type at family carries out pattern match at the class cluster center of cluster, to carry out diversity project for the active user
Information recommendation, comprising:
According to the user type of the active user in the following two kinds interest preference mode at the class cluster center of cluster into
Row pattern match: regular interest preference mode and non-regularity interest preference mode；Wherein, the regular interest preference mould
Formula is to like recently and also like in the past；The non-regularity interest preference mode includes: to like recently but do not like in the past, most
It closely do not like but likes in the past, do not like recently and do not like in the past yet.
5. the personalized recommendation method as described in claim 1 based on time entropy, which is characterized in that described according to the current use
The association attributes characteristic information at family and user calculate the different scoring records of project using following time entropy formula described current
User is to item object time entropy:
Wherein, H is time entropy, and n is time interval number, scoreiScoring for user to project i, miRepresent a time interval
Interior all users refer to all m to the scoring summation of project i, MiSummation.
6. a kind of personalized recommendation device based on time entropy, which is characterized in that described device includes:
Cluster cell obtains the class cluster center of cluster and each for clustering to existing user and project score information
The user for including in class cluster and project information；
Computing unit, for if it is determined that active user is not belonging to that system user newly is added, then according to the correlation of the active user
Attributive character information and user calculate the active user to item object time entropy to the different scoring records of project；
Judging unit, for, to item object time entropy, determining the use of the active user according to the calculated active user
Family type；
Matching unit, for carrying out pattern match at the class cluster center of cluster according to the user type of the active user,
To carry out the recommendation of diversity project information for the active user.
7. as claimed in claim 6 based on the personalized recommendation device of time entropy, which is characterized in that described device further include:
Recommendation unit, be used for if it is determined that active user be it is new system user is added, will using following project popularity calculation
The preceding N number of project information of ranking recommends the active user:
Wherein, ipopFor project popularity, UiUser's set that project i was evaluated in recommender system is represented, I is in recommender system
All items number.
8. as claimed in claim 6 based on the personalized recommendation device of time entropy, which is characterized in that
The judging unit is specifically used for determining according to user interest preference that the user type of the active user is following four
One of seed type: liking recently and also likes in the past, likes recently but do not like in the past, do not like recently but like in the past, most
It does not like closely and does not like in the past yet.
9. as claimed in claim 8 based on the personalized recommendation device of time entropy, which is characterized in that
The matching unit, specifically for the user type according to the active user in the as follows of the class cluster center of cluster
Pattern match is carried out in two kinds of interest preference modes: regular interest preference mode and non-regularity interest preference mode；Wherein,
The regularity interest preference mode is to like recently and also like in the past；The non-regularity interest preference mode includes: most
Closely like but do not like in the past, do not like recently but like in the past, do not like recently and does not also like in the past.
10. as claimed in claim 6 based on the personalized recommendation device of time entropy, which is characterized in that
The computing unit, specifically for the difference according to the association attributes characteristic information of the active user and user to project
Scoring record calculates the active user to item object time entropy using following time entropy formula:
Wherein, H is time entropy, and n is time interval number, scoreiScoring for user to project i, miRepresent a time interval
Interior all users refer to all m to the scoring summation of project i, MiSummation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810794746.2A CN109063052B (en) | 2018-07-19 | 2018-07-19 | Personalized recommendation method and device based on time entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810794746.2A CN109063052B (en) | 2018-07-19 | 2018-07-19 | Personalized recommendation method and device based on time entropy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109063052A true CN109063052A (en) | 2018-12-21 |
| CN109063052B CN109063052B (en) | 2022-01-25 |
Family
ID=64817411
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810794746.2A Expired - Fee Related CN109063052B (en) | 2018-07-19 | 2018-07-19 | Personalized recommendation method and device based on time entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109063052B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110070134A (en) * | 2019-04-25 | 2019-07-30 | 厦门快商通信息咨询有限公司 | A kind of recommended method and device based on user interest perception |
| CN111191707A (en) * | 2019-12-25 | 2020-05-22 | 浙江工商大学 | LFM training sample construction method fusing time attenuation factors |
| CN114548501A (en) * | 2022-01-14 | 2022-05-27 | 北京全路通信信号研究设计院集团有限公司 | Balance checking method, system and equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101329683A (en) * | 2008-07-25 | 2008-12-24 | 华为技术有限公司 | Recommendation system and method |
| US20100185579A1 (en) * | 2009-01-22 | 2010-07-22 | Kwang Seok Hong | User-based collaborative filtering recommendation system and method for amending similarity using information entropy |
| CN102135999A (en) * | 2011-03-25 | 2011-07-27 | 南京财经大学 | User credibility and item nearest neighbor combination Internet recommendation method |
| CN103186539A (en) * | 2011-12-27 | 2013-07-03 | 阿里巴巴集团控股有限公司 | Method and system for confirming user groups, inquiring information and recommending |
| CN106649658A (en) * | 2016-12-13 | 2017-05-10 | 重庆邮电大学 | Recommendation system and method for improving user role undifferentiated treatment and data sparseness |
| CN107391713A (en) * | 2017-07-29 | 2017-11-24 | 内蒙古工业大学 | A kind of method and system for solving the problems, such as cold start-up in collaborative filtering recommending technology |
-
2018
- 2018-07-19 CN CN201810794746.2A patent/CN109063052B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101329683A (en) * | 2008-07-25 | 2008-12-24 | 华为技术有限公司 | Recommendation system and method |
| US20100185579A1 (en) * | 2009-01-22 | 2010-07-22 | Kwang Seok Hong | User-based collaborative filtering recommendation system and method for amending similarity using information entropy |
| CN102135999A (en) * | 2011-03-25 | 2011-07-27 | 南京财经大学 | User credibility and item nearest neighbor combination Internet recommendation method |
| CN103186539A (en) * | 2011-12-27 | 2013-07-03 | 阿里巴巴集团控股有限公司 | Method and system for confirming user groups, inquiring information and recommending |
| CN106649658A (en) * | 2016-12-13 | 2017-05-10 | 重庆邮电大学 | Recommendation system and method for improving user role undifferentiated treatment and data sparseness |
| CN107391713A (en) * | 2017-07-29 | 2017-11-24 | 内蒙古工业大学 | A kind of method and system for solving the problems, such as cold start-up in collaborative filtering recommending technology |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110070134A (en) * | 2019-04-25 | 2019-07-30 | 厦门快商通信息咨询有限公司 | A kind of recommended method and device based on user interest perception |
| CN111191707A (en) * | 2019-12-25 | 2020-05-22 | 浙江工商大学 | LFM training sample construction method fusing time attenuation factors |
| CN114548501A (en) * | 2022-01-14 | 2022-05-27 | 北京全路通信信号研究设计院集团有限公司 | Balance checking method, system and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109063052B (en) | 2022-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11645301B2 (en) | Cross media recommendation | |
| Li et al. | Using multidimensional clustering based collaborative filtering approach improving recommendation diversity | |
| CN105224699B (en) | News recommendation method and device | |
| Qian et al. | Personalized recommendation combining user interest and social circle | |
| TWI636416B (en) | Method and system for multi-phase ranking for content personalization | |
| US9405746B2 (en) | User behavior models based on source domain | |
| Liu et al. | Social temporal collaborative ranking for context aware movie recommendation | |
| CN104969224B (en) | Improved user experience for unacknowledged and new users | |
| CN103678518B (en) | Method and device for adjusting recommendation lists | |
| CN109241405A (en) | A kind of associated education resource collaborative filtering recommending method of knowledge based and system | |
| CN105574045B (en) | Video recommendation method and server | |
| KR101098871B1 (en) | APPARATUS AND METHOD FOR MEASURING CONTENTS SIMILARITY BASED ON FEEDBACK INFORMATION OF RANKED USER and Computer Readable Recording Medium Storing Program thereof | |
| CN109511015B (en) | Multimedia resource recommendation method, device, storage medium and equipment | |
| CN108540826A (en) | Barrage method for pushing, device, electronic equipment and storage medium | |
| Boratto et al. | The rating prediction task in a group recommender system that automatically detects groups: architectures, algorithms, and performance evaluation | |
| Hu et al. | Style in the long tail: Discovering unique interests with latent variable models in large scale social e-commerce | |
| Yazdanfar et al. | Link recommender: Collaborative-filtering for recommending urls to twitter users | |
| KR102108683B1 (en) | Method for providing recommendation contents including non-interest contents | |
| CN109063052A (en) | A kind of personalized recommendation method and device based on time entropy | |
| WO2023231542A1 (en) | Representation information determination method and apparatus, and device and storage medium | |
| KR20140015653A (en) | Contents recommendation system and contents recommendation method | |
| KR102164409B1 (en) | Crowdsourcing user-provided identifiers and associating them with brand identities | |
| CN111078997A (en) | Information recommendation method and device | |
| KR20170060828A (en) | User's Contents Access History based Recommendation Method | |
| CN116186119A (en) | User behavior analysis method, device, equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220125 |