CN103399919A - Trust enhanced service push method based on social relation network - Google Patents
Trust enhanced service push method based on social relation network Download PDFInfo
- Publication number
- CN103399919A CN103399919A CN2013103332451A CN201310333245A CN103399919A CN 103399919 A CN103399919 A CN 103399919A CN 2013103332451 A CN2013103332451 A CN 2013103332451A CN 201310333245 A CN201310333245 A CN 201310333245A CN 103399919 A CN103399919 A CN 103399919A
- Authority
- CN
- China
- Prior art keywords
- service
- user
- trust
- receipt
- migration
- 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
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a service push method based on a trust network. The service push method includes the steps of performing a trust weighted operation on a social relation network between users, assessing the trust degree between the directly-connected users, building a weighted trust network, obtaining prediction scores of services which are not called by a target user through a trust enhanced random walk method, determining the sequence of push services according to the service prediction scores graded by the target user, and pushing a specific number of push services to the target user sequentially from high to low. According to the method, pushing is performed while the trust degree and the similarity between the users are considered, nodes similar to the target user tend to be selected in each walking process, and pushing efficiency is improved effectively.
Description
Technical field
The present invention relates to Web data-pushing technical field, relate in particular to a kind of service push method based on trust network.
Background technology
Software service take web services as representative and software service are worked in coordination with has become a kind of novel web application form.Along with the web services that is deployed on Internet is enriched constantly, these can have been formed a huge standard package storehouse by public visit and integrated service.But in the face of the identical or close web services of a large amount of functions, the user is at a loss as to what to do on services selection, and the user is difficult to be identified to the quality requirements of service, only meets simply the response time or reduces the preference demand that expense can not meet each user.So, very important according to the services selection index of user's preference customized personal.Along with increasing user accesses the formed experience accumulation of web services, how according to user's request, numerous candidate service being selected and it is pushed to the user has become an important research direction.
In recent years, along with Web2.0 develops rapidly, internet, from the stationary platforms of a propagation and sharing information, has become the social interaction platform that participates in more widely, creates, links up.Various types of online community networks (onlinesocial network) emerge like the mushrooms after rain, such as Myspace, Facebook, Twitter etc. of several famous social network sites successfully ranked among the row of the website that the global access amount is the highest, becomes 2l century one of the most significant technical phenomena.And the dynamic that the Web service environment has and opening, and participant's independence makes itself and the environment of social network quite similar, its similarity is mainly manifested in: 1) service consumer has autonomous right to choose fully for interactive object; 2) interactive information can be shared; 3) can collect, analyze about service interactive information decision in the past whether set up interactive relation; 4) between the user, can utilize the transmission of informations such as trust delegation, propelling movement; 5) serve that the participant is obligated provides pushed information for other individualities under the web services environment.Therefore, this provides possibility for some conclusion of using for reference The Study of Sociology.Under the ordering about of community network, user's wide participation makes the description of service except functional attributes and NOT-function attribute description that the ISP provides, also there is user's feedback information, comprise that the user uses history, user to the evaluation of specific service and the information such as social relationships between the user, utilize these information can excavate the user to the personalization preferences of service, trusting relationship, user between the user to other users' influence power etc., thereby more accurately, for the user, carry out the personalized service propelling movement objectively.As can be seen here, rationally excavate the information in community network, can bring new opportunity and challenge for realizing that more effectively service personalization pushes.
At present, the research that pushes for web services mainly concentrates on service quality (QoS, the Quality-of-Service) prediction to web services, according to the qos value sequence that prediction draws, the user is served to propelling movement.But this class method for pushing is often only paid close attention to the attribute with service self, and has ignored the direct feel of user to service, is difficult to the personalization preferences of abundant digging user.And, along with the introducing of social networks, will collect more field feedback, can utilize these information to carry out the personalized service propelling movement.Supplying system based on field feedback extensively is present in the propelling movement websites such as Amazon, Epinions, Netflix.Utilize collaborative filtering to push algorithm according to the expressed feedback opinion that goes out of user etc., search the similar neighbours of user or project, thereby produce, push result.But these class methods exist certain drawback, 1) the cold start-up problem, for the user who there is no historical feedback information, can't push; 2) sparse property problem, when user's feedback information is too sparse, can't guarantee higher propelling movement accuracy; 3) Dependability Problem, can exist some fictitious users feedback informations to cause interference to pushing result.
For the problems referred to above, a few thing has proposed to trust the propelling movement that strengthens, and these class methods, according to user's direct or indirect trust user, are predicted its scoring to project, thereby pushed, and for solving above-mentioned three problems, provide new thinking.Present existing trust strengthens method for pushing often based on 0/1 users to trust relation, as long as the user that target is trusted can push as its neighbor user.The method thinks that all scorings from trusted users are all of equal value on the impact of prediction scoring, but in fact, due to the preference from the targeted customer and to different perception degree of trusting users, there are differences, for difference trust user's scoring, adopt degree and also should distinguish to some extent.
Summary of the invention
The present invention is directed to the method for pushing deficiency that existing trust strengthens, a kind of service push method based on trust network is provided, the method can, according to targeted customer's preference, be adopted degree for difference trust user's scoring and distinguish.
A kind of enhancement service of trust based on social relationships net method for pushing comprises: 1) calculate the trust degree of correlation between the user, set up weighting and trust the customer relationship net, wherein, the described trust degree of correlation is according to formula: tr (u, v)=SimU (u, v) * t (u, v) calculate the trust degree of correlation similarity of tr (u, v) expression user u and user v, t (u, v) be the degree of belief of user u to user v, the similarity of simU (u, v) expression user u and user v;
2) by trust, strengthen the scoring of random walk method target of prediction user to the service of never call, take the targeted customer as starting point, in described weighting, trust in the customer relationship net and carry out time migration of many receipt, using the mean value of many receipt time migration result as final score in predicting result;
3) according to the targeted customer, to the prediction scoring of service, determine the order of Push Service, to the targeted customer, push successively from high to low the Push Service of specified quantity.
The present invention trusts the weighting operation by the social relationships net between the user, trusting degree between the user that i.e. assessment directly is connected, and serve propelling movement by random roaming algorithm, considered that simultaneously degree of belief and similarity between the user push, while making each migration, more trend towards selecting the node similar to the targeted customer, thereby improve the efficiency that pushes.
Described
And as cos (u`, v`) less than 0 the time, simU (u, v)=0, u`, v` are respectively the user characteristics vector of family u and user v.
Cos (u`, v`) span is [1,1], but between its value is less than two users of 0 time explanation for the preference impression of service the restricted publication of international news and commentary entitled value too not during for Push Service with, this class user hardly, therefore as cos (u`, v`) less than 0 o'clock, the value of cos (u`, v`) is 0 by simU (u, v) span is limited between [0,1].
Described step 2) in, trusting enhancing random walk method is: set the maximum migration distance of single migration, in the scope of this maximum migration distance, node of every arrival is all done as judged:
If the user u that a) this node is corresponding comments on, treated Push Service s, and directly user u was treated to the scoring of Push Service s and return as the result of this receipt time migration, time migration of this receipt stops, otherwise execution step b);
B) calculate the probability that stops of migration
If meeting imposes a condition, stop at present node u, the set of service RS that then estimated from user u corresponding to present node u
uMiddle according to selecting probability F
u(si) select with reference to Push Service s
i, and by user u to reference Push Service s
iScoring as the returning results of this receipt time migration, time migration of this receipt stops, if stop probability
Do not meet and impose a condition, select and enter next node, proceed time migration of this receipt;
C) if time migration of this receipt, to maximum migration distance, does not meet step a), b yet) in stop condition, stop at final node u
L, then from rearmost point u
LCorresponding user u
LThe set of service RS that estimated
uMiddle according to selecting probability F
u(si) select with reference to Push Service s
i, and by user u
LTo reference Push Service s
iScoring as the returning results of this receipt time migration, time migration of this receipt stops, and carries out time migration of next receipt.
After time migration of every receipt stops, judging whether to satisfy condition | σ
I+1 2-σ
i 2|≤ε, ε=0.0001, wherein σ
I+1 2Represent σ
i 2Be respectively front i+1 return with time migration of front i receipt after the variance that returns results,
r
jReturning results of j receipt time migration,
The mean value that before expression, time migration of i receipt returns results, if meet, whole trust enhancing random walk finishes; Otherwise, get back to the targeted customer, proceed time migration of next receipt.
Trust the enhancing random walk and comprise time migration of many receipt, a result is returned in time migration of every receipt.By time migration of many receipt, by the results averaged of the many receipt time migration prediction appraisal result of targeted customer to current service the most, the result that obtains is more accurate.In single migration process, time migration of every receipt all judges that time migration of this receipt is to judge whether to meet stop condition according to present node during to a certain node, while meeting, according to this node, obtains an appraisal result and returns, and time migration of this receipt stops; While not meeting, select the node of next migration, proceed time migration of this receipt, until meet stop condition, according to the node that stops, obtaining an appraisal result and return, time migration of this receipt stops.Due to random walk, there is the possibility that forever can't stop, therefore setting the maximum step-length of single random walk.
In single migration process, for present node u, select the probability of next migration node v to be:
TU
uUser's set that expression user u directly trusts, x represents the user in this set.
By this restrictive condition, obtain each migration and more trend towards selecting the trust user more similar to active user's interest preference, thereby make the reference significance of its propelling movement larger, reduced process step-length in the single migration, help to improve migration efficiency.
The described probability that stops
Impose a condition and with reference to Push Service s
iMeet following relation:
The probability that in the expression random walk, the k step stops at user u, simS (s
i, s) expression is with reference to Push Service s
iWith the similarity for the treatment of Push Service s.
The service of estimating with user u is relevant with the similarity for the treatment of Push Service s, its service of estimating to treat that Push Service s is more similar, the probability that stops is just larger.In addition, user u and targeted customer u
0Between distance far away, the interference that predicts the outcome is just larger.Therefore, along with the increase of k value,
Also should increase.
Described
And as cos (s
i`, s
j`) less than 0 o'clock, simS (s
i, s
j)=0, s
i`, s` are respectively s
iWith the service characteristic vector for the treatment of Push Service s.
In prior art, mostly adopt project-based collaborative filtering method to calculate the similarity between two services, the drawback of this method is, in case there is not the user of common evaluation in two services, so just can't calculate their similarity value.Cosine method for measuring similarity according to vector in the present invention calculates the similarity between two users, two services, does not have the user of common evaluation, still can obtain their similarity value.
Described probability
Described user characteristics proper vector and service are levied proper vector and are obtained by lower method:
1-1) set up the user of m * n-service rating matrix R, m is number of users, and n is the service number;
1-2) user that sets up-service rating matrix R is carried out to dimensionality reduction, employing Singular Value Decomposition Using technology, be decomposed into the eigenmatrix S of user characteristics matrix U and service by user-service rating matrix R, meets R ≈ US
T, the user characteristics vector of the user of each line display in U, the service characteristic vector of the service of each line display in S.
In existing service supplying system, the user is very huge with the rating matrix that service forms, and therefore the user is recorded as proper vector the scoring of service, and its dimension is quite high, and wherein a lot of score data lack.The present invention decomposes user-service rating matrix by Singular Value Decomposition Using technology (SVD) and obtains simultaneously user characteristics vector sum service characteristic vector, and the user characteristics vector sum service characteristic vector dimension that obtains is low, has effectively reduced the operand of system.
Trust based on social relationships net enhancement service method for pushing provided by the invention.At first the method trusts the weighting operation to the social relationships net between the user, i.e. trusting degree between the user that assessment directly is connected.Then, propose a kind of improved random roaming algorithm and served propelling movement, this algorithm has considered that simultaneously degree of belief and the similarity between the user pushes, and while making each migration, more trends towards selecting the node similar to the targeted customer, thereby improves the efficiency that pushes.
The accompanying drawing explanation
Fig. 1 is the process flow diagram of the enhancement service of the trust based on the social relationships net method for pushing of the present embodiment.
Embodiment
The present invention is described further below in conjunction with specific embodiment.
Known users collection U={u
1, u
2... u
n, services set S={s
1, s
2... s
n.Each user u can be to one group of service S
u={ s
U1, s
U2... s
UnMark, user u is shown r to the grade form of service s
u,s, general integer representation with [1,5].Between the user, according to trusting relationship each other, set up a social relation network.If user u trusts user v, t (u, v) can be expressed as the degree of belief of user u to user v, adopts 0/1 trusting relationship, makes according to the trusting relationship between the user that the value of t (u, v) is 0 or 1, TU
u={ v ∈ U|t (u, v)=1} represents user's set that user u directly trusts.Trust network can be expressed as a digraph SN=<U, T>, T={ (u, v) wherein | u ∈ U, v ∈ T
u, each node is corresponding to a user, and every limit is corresponding to a trusting relationship.In the present embodiment, the targeted customer is u
0, u
0∈ U, treat the trust network SN that forms between Push Service s ∈ S and user,
The unknown, prediction u
0Scoring to service s.
Definition such as the table 1 of each variable in the present embodiment:
Table 1
The flow process of the enhancement service of the trust based on the social relationships net method for pushing of the present embodiment as shown in Figure 1, comprises the following steps:
1) according to formula: tr (u, v)=SimU (u, v) * t (u, v) calculate the trust degree of correlation between the user, set up weighting and trust the customer relationship net, wherein, tr (u, v) the trust degree of correlation similarity of expression user u and user v, t (u, v) is the degree of belief of user u to user v, simU (u, v) similarity of expression user u and user v
And as cos (u`, v`) less than 0 the time, simU (u, v)=0, u`, v` are respectively the user characteristics vector of family u and user v.
2) by trust, strengthen random walk, the scoring of target of prediction user to the service of never call, trust enhancing random walk is included in weighting trust customer relationship net carries out time migration of many receipt, using the mean value of many receipt time migration result as final score in predicting result, the scoring of target of prediction user to the service of never call.
The present invention introduces the concept of " Six Degrees ", and the maximum migration distance of setting the single migration is 6, and each migration is all with targeted customer u
0For starting point, in the scope of this maximum migration distance, node of every arrival is all done as judged:
If the user u that a) this node is corresponding comments on, treated Push Service s, directly user u was treated to the scoring of Push Service s as this
After time migration of every receipt stops, judging whether to satisfy condition | σ
I+1 2-σ
i 2|≤ε, ε=0.0001, wherein σ
I+1 2Represent σ
i 2Be respectively front i+1 return with time migration of front i receipt after the variance that returns results,
r
jReturning results of j receipt time migration,
The mean value that before expression, time migration of i receipt returns results, if do not meet, get back to targeted customer u
0, proceed time migration of next receipt, otherwise this trusts enhancing random walk end.This is trusted to many receipt time the returning results of migration that strengthens in the random walk process averages as user u
0Prediction scoring to service s
Namely
N trusts total number that returns of the single migration that strengthens in random walk for this.
In every receipt time migration process:
For present node u, select the probability of next migration node v to be:
Select the probability F of service si
u(s
i):
Wherein, simS (s
i, s) expression is with reference to recommendation service s
iWith the similarity of service s to be recommended, as cos (s
i`, s
j`) less than 0 o'clock, simS (s
i, s
j)=0, otherwise,
s
i`, s` are respectively with reference to recommendation service s
iService characteristic vector with service s to be recommended.
User characteristics proper vector and service are levied proper vector and are obtained by lower method:
1-1) set up the user of m * n-service rating matrix R, m is number of users, and n is the service number;
1-2) user that sets up-service rating matrix R being carried out to dimensionality reduction is the d dimension, adopts the Singular Value Decomposition Using technology, and user-service rating matrix R is decomposed into to the user characteristics matrix U and serves eigenmatrix S, meets R ≈ US
T, user characteristics matrix U ∈ R
M * dEigenmatrix S ∈ R with service
N * d, the user characteristics vector of the user of each line display in the user characteristics matrix U, the service characteristic vector of the service of each line display in the service features matrix S.
Repeatedly trust the enhancing random walk, obtain the prediction scoring of targeted customer to a plurality of services, according to the prediction scoring of targeted customer to service, determine the order of Push Service, by the prediction scoring, to the targeted customer, push 10 services successively from high to low.
Claims (9)
1. the enhancement service of the trust based on a social relationships net method for pushing, is characterized in that, comprising:
1) calculate the trust degree of correlation between the user, set up weighting and trust the customer relationship net, wherein, the described trust degree of correlation is according to formula: tr (u, v)=SimU (u, v) * t (u, v) calculates, tr (u, v) the trust degree of correlation similarity of expression user u and user v, t (u, v) is the degree of belief of user u to user v, the similarity of simU (u, v) expression user u and user v;
2) by trust, strengthen the scoring of random walk method target of prediction user to the service of never call, take the targeted customer as starting point, in described weighting, trust in the customer relationship net and carry out time migration of many receipt, using the mean value of many receipt time migration result as final score in predicting result;
3) according to the targeted customer, to the prediction scoring of service, determine the order of Push Service, to the targeted customer, push successively from high to low the Push Service of specified quantity.
2. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 1, is characterized in that, and is described
And as cos (u`, v`) less than 0 the time, simU (u, v)=0, u`, v` are respectively the user characteristics vector of family u and user v.
3. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 2, it is characterized in that, described step 2) in, trusting enhancing random walk method is: the maximum migration distance of setting the single migration, in the scope of this maximum migration distance, node of every arrival is all done as judged:
If the user u that a) this node is corresponding comments on, treated Push Service s, and directly user u was treated to the scoring of Push Service s and return as the result of this receipt time migration, time migration of this receipt stops, otherwise execution step b);
B) calculate the probability that stops of migration
If meeting imposes a condition, stop at present node u, the set of service RS that then estimated from user u corresponding to present node u
uMiddle according to selecting probability F
u(s
i) select with reference to Push Service s
i, and by user u to reference Push Service s
iScoring as the returning results of this receipt time migration, time migration of this receipt stops, if stop probability
Do not meet and impose a condition, select and enter next node, proceed time migration of this receipt;
C) if time migration of this receipt, to maximum migration distance, does not meet step a), b yet) in stop condition, stop at final node u
L, then from rearmost point u
LCorresponding user u
LThe set of service RS that estimated
uMiddle according to selecting probability F
u(s
i) select with reference to Push Service s
i, and by user u
LTo reference Push Service s
iScoring as the returning results of this receipt time migration, time migration of this receipt stops, and carries out time migration of next receipt.
4. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 3, is characterized in that, after time migration of every receipt stops, judging whether to satisfy condition | σ
I+1 2-σ
i 2|≤ε, ε=0.0001, wherein σ
I+1 2Represent σ
i 2Be respectively front i+1 return with time migration of front i receipt after the variance that returns results,
r
jReturning results of j receipt time migration,
The mean value that before expression, time migration of i receipt returns results, if meet, whole trust enhancing random walk finishes; Otherwise, get back to the targeted customer, proceed time migration of next receipt.
5. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 4, is characterized in that, in single migration process, for present node u, selects the probability of next migration node v to be:
TU
uUser's set that expression user u directly trusts, x represents the user in this set.
6. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 5, is characterized in that the described probability that stops
Impose a condition and with reference to Push Service s
iMeet following relation:
7. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 6, is characterized in that, and is described
And as cos (s
i`, s
j`) less than 0 o'clock, simS (s
i, s
j)=0, s
i`, s` are respectively with reference to Push Service s
iWith the service characteristic vector for the treatment of Push Service s.
9. the enhancement service of the trust based on social relationships net method for pushing as claimed in claim 8, is characterized in that, described user characteristics proper vector and service are levied proper vector and obtained by lower method:
1-1) set up the user of m * n-service rating matrix R, m is number of users, and n is the service number;
1-2) user that sets up-service rating matrix R is carried out to dimensionality reduction, employing Singular Value Decomposition Using technology, be decomposed into the eigenmatrix S of user characteristics matrix U and service by user-service rating matrix R, meets R ≈ US
T, the user characteristics vector of the user of each line display in U, the service characteristic vector of the service of each line display in S.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103332451A CN103399919A (en) | 2013-08-02 | 2013-08-02 | Trust enhanced service push method based on social relation network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103332451A CN103399919A (en) | 2013-08-02 | 2013-08-02 | Trust enhanced service push method based on social relation network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103399919A true CN103399919A (en) | 2013-11-20 |
Family
ID=49563547
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2013103332451A Pending CN103399919A (en) | 2013-08-02 | 2013-08-02 | Trust enhanced service push method based on social relation network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103399919A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105045805A (en) * | 2015-06-03 | 2015-11-11 | 安徽师范大学 | Construction method for trust recommendation model based on mobile communications and construction system thereof |
CN105354260A (en) * | 2015-10-22 | 2016-02-24 | 中南大学 | Mobile application recommendation method with social network and project feature fused |
CN110175296A (en) * | 2018-11-08 | 2019-08-27 | 腾讯科技(深圳)有限公司 | Nodes recommendations method and server and storage medium in network |
CN111371619A (en) * | 2020-03-10 | 2020-07-03 | 广州大学 | Method and system for estimating number of users of instant messaging network |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080109451A1 (en) * | 2006-10-17 | 2008-05-08 | Harding Benjamin L | Method and system for evaluating trustworthiness |
CN102364527A (en) * | 2011-10-21 | 2012-02-29 | 中国科学技术大学 | Real-time identity recognition and authentication method for self-service equipment system of bank |
CN102880640A (en) * | 2012-08-20 | 2013-01-16 | 浙江大学 | Network modeling-based service recommending method |
-
2013
- 2013-08-02 CN CN2013103332451A patent/CN103399919A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080109451A1 (en) * | 2006-10-17 | 2008-05-08 | Harding Benjamin L | Method and system for evaluating trustworthiness |
CN102364527A (en) * | 2011-10-21 | 2012-02-29 | 中国科学技术大学 | Real-time identity recognition and authentication method for self-service equipment system of bank |
CN102880640A (en) * | 2012-08-20 | 2013-01-16 | 浙江大学 | Network modeling-based service recommending method |
Non-Patent Citations (3)
Title |
---|
王丽莎: "基于随机游走模型的个性化信息推荐", 《中国优秀硕士学位论文全文数据库科技信息辑》, no. 7, 15 July 2012 (2012-07-15) * |
顾晔: "改进的增量奇异值分解协同过滤算法", 《计算机工程与应用》 * |
高佳祥: "基于加权网络的协同过滤推荐", 《中国优秀硕士学位论文全文数据库科技信息辑》, no. 1, 15 January 2013 (2013-01-15), pages 22 - 42 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105045805A (en) * | 2015-06-03 | 2015-11-11 | 安徽师范大学 | Construction method for trust recommendation model based on mobile communications and construction system thereof |
CN105045805B (en) * | 2015-06-03 | 2018-07-24 | 安徽师范大学 | The construction method and its structure system of trust recommendation model based on mobile communication |
CN105354260A (en) * | 2015-10-22 | 2016-02-24 | 中南大学 | Mobile application recommendation method with social network and project feature fused |
CN110175296A (en) * | 2018-11-08 | 2019-08-27 | 腾讯科技(深圳)有限公司 | Nodes recommendations method and server and storage medium in network |
CN110175296B (en) * | 2018-11-08 | 2023-03-24 | 腾讯科技(深圳)有限公司 | Node recommendation method and server in network graph and storage medium |
CN111371619A (en) * | 2020-03-10 | 2020-07-03 | 广州大学 | Method and system for estimating number of users of instant messaging network |
CN111371619B (en) * | 2020-03-10 | 2022-06-10 | 广州大学 | Method and system for estimating number of users of instant messaging network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11659050B2 (en) | Discovering signature of electronic social networks | |
CN107590243B (en) | The personalized service recommendation method to be sorted based on random walk and diversity figure | |
Pham et al. | A general graph-based model for recommendation in event-based social networks | |
Symeonidis et al. | Geo-social recommendations based on incremental tensor reduction and local path traversal | |
Li et al. | A dynamic algorithm based on cohesive entropy for influence maximization in social networks | |
CN103473291A (en) | Personalized service recommendation system and method based on latent semantic probability models | |
CN103024017A (en) | Method for distinguishing important goals and community groups of social network | |
CN108334575A (en) | A kind of recommendation results sequence modification method and device, electronic equipment | |
CN105809554A (en) | Prediction method of hot topics participated by users in social networks | |
CN106951471A (en) | A kind of construction method of the label prediction of the development trend model based on SVM | |
US20140229510A1 (en) | Method of temporal bipartite projection | |
CN103399919A (en) | Trust enhanced service push method based on social relation network | |
CN103198432B (en) | Detection method and detection system of network groups in online social network | |
CN116244513A (en) | Random group POI recommendation method, system, equipment and storage medium | |
Postigo-Boix et al. | A social model based on customers’ profiles for analyzing the churning process in the mobile market of data plans | |
Xiong et al. | Reward optimization for content providers with mobile data subsidization: A hierarchical game approach | |
Hori et al. | Inflation expectations of Japanese households: Micro evidence from a Consumer Confidence Survey | |
CN112800345A (en) | Community role-aware user demand active prediction method and system | |
Papadimitriou et al. | Geo-social recommendations | |
Yang et al. | A game theoretic model for the formation of navigable small-world networks | |
Tibermacine et al. | Regression-based bootstrapping of web service reputation measurement | |
Jia et al. | TSBM: The temporal-spatial Bayesian model for location prediction in social networks | |
Zhang | Research on collaborative filtering recommendation algorithm based on social network | |
CN102521362B (en) | Web service recommendation method and device | |
CN111143700A (en) | Activity recommendation method and device, server and computer storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131120 |
|
RJ01 | Rejection of invention patent application after publication |