CN105791085B - Friend recommendation method in the social networks of position based on position and time - Google Patents
Friend recommendation method in the social networks of position based on position and time Download PDFInfo
- Publication number
- CN105791085B CN105791085B CN201610051508.3A CN201610051508A CN105791085B CN 105791085 B CN105791085 B CN 105791085B CN 201610051508 A CN201610051508 A CN 201610051508A CN 105791085 B CN105791085 B CN 105791085B
- Authority
- CN
- China
- Prior art keywords
- user
- place
- classification
- registering
- information
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L51/00—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
- H04L51/52—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail for supporting social networking services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Abstract
The friend recommendation method based on position and time that the invention discloses a kind of mainly solves the problems, such as that existing friend recommendation scheme ignores the temporal information in information of registering and position semantic information of registering causes to recommend precision not high enough.Implementation step is：1. establishing the communication system frame being made of user, location facilities, location server；2. the longitude and latitude of position and current temporal information are sent to social network server and generate information of registering by user；3. social network server pre-processes the information of largely registering of storage；4. social network server calculates each user's term frequency-inverse document frequency score other for each location category；5. social network server calculates the similarity between user, and chooses the higher user of similarity and make friend recommendation.The present invention alleviates recommender system data available sparsity, improves recommendation accuracy.It can be used for location-based wireless social networking service.
Description
Technical field：
The invention belongs to radio network technique fields, are related to the friend recommendation of wireless social networks, can be applied to based on position
The wireless social networking service set.
Background technique：
Location-based social networks helps user to share their real time position online in order to which user can be found that sense
It simultaneously makes friends the position of interest.For example, user can have found oneself interested place by the position that good friend shares, or
New good friend is made friends with by searching for the user in similar place is shared with oneself.In addition, popularizing with smart phone, built in
GPS module can be more accurate detection user position, allow user more easily to share respective positions.Cause
This, this service of registering has attracted more and more users.It how the use of information of largely registering to be that user carries out friend recommendation
It needs to be paid attention to.
Recommender system plays important role in social networks and e-commerce website, in e-commerce website,
Existing recommender system analyzes the preference of user usually using the purchaser record of user, as the purchase of e-commerce website is remembered
Record, register history of the user in location-based social networks also contain the largely information about user preference.
When carrying out friend recommendation using the information of registering of user, since user usually only can a small portion in the database
Point place carries out behavior of registering, therefore the data available for carrying out friend recommendation is more sparse.The problem of in order to alleviate Sparse,
Existing scheme is divided into following several：Classified according to the density in place of registering, is registered what some geographical locations were closer to
Regard the same place of registering as in place；It is filled into the data of registering of user using the data of registering of user good friend.But it is geographical
The place being closely located to can represent the different hobby of user, and equally, the hobby of good friend can not replace user completely
The hobby of itself, both behaviors all reduce the precision of friend recommendation to a certain extent.
In terms of the recommender system for considering temporal information, existing system is mostly used to recommend net in e-commerce website and film
It stands, the preference of user is recorded and analyzed by the viewing of analysis user long-term purchaser record and film to recommend.Compared to
The behavior of registering of long-term trend, user can provide the information more to reflect user preferences in one day specific time.And
In existing social networks recommender system, only place of registering is classified by analyzing the time attribute in place of registering, not
The register Annual distribution of behavior of user is analyzed and carries out friend recommendation, therefore the accuracy rate recommended is not ideal enough.
Summary of the invention
It is an object of the invention to be directed to the deficiency of above-mentioned prior art, propose one kind in the social networks of position based on position
The friend recommendation method with the time is set, to improve the accuracy rate recommended.
The present invention, which passes through, is classified as one kind for the place for possessing similar semantic information, and it is sparse to solve data available in recommender system
The problem of, it registers row using the interest in term frequency-inverse document frequency equilibrium hot topic place and user itself, while by analyzing user
For the regularity of distribution at any time, the similarity calculation of user is adjusted, realizes more accurate friend recommendation.Its realization side
Case includes as follows：
(1) the communication system frame being made of user, location facilities, social network server is established, wherein：
User, for being communicated by mobile cellular network or WiFi with location facilities and social network server；
Location facilities realize positioning for collaborative user's GPS module；
Social network server carries out friend recommendation to user for storing the location information of user, and on this basis；
(2) longitude and latitude of position and current temporal information are sent to social network server by user, generate letter of registering
Breath；
(3) social network server pre-processes the information of registering of storage：
(3a) is directed to the longitude and latitude in information of registering, and the position semantic database for searching server obtains and place pair of registering
Then the information of registering of each user is counted number of registering by the semantic information classification in its place of registering by the semantic information answered；
The temporal information that (3b) is directed in information of registering counts each user eachly on the basis of step (3a)
The distribution of behavior at any time of registering under point classification；
(4) according to user in the number of registering of each place classification, each user u of social network server calculating is for each
The term frequency-inverse document frequency score TF-IDF of place classification cu,c；
The number of registering of (4a) according to user u under each place classification, social network server calculate user u eachly
The word frequency score value TF for information of registering under point classification cu,c；
The number of registering of (4b) according to all users under each place classification, social network server calculate each location category
The inverse document frequency score value IDF of other cc；
Word frequency score value TF of (4c) social network server by user u at the classification c of placeu,cWith the inverse text of place classification c
Shelves frequency score IDFcIt is multiplied, obtains its term frequency-inverse document frequency score TF-IDF under the place classificationu,c=TFu,c×
IDFc；
(5) social network server calculates the similarity between user：
(5a) is registered the distribution of behavior at any time using user under each place classification, calculates two users in the same manner
The relative entropy D (P Q) for distribution of registering under classification is put, wherein P and Q is two user uiAnd ujIt is secondary in registering for identical place classification
The probability distribution of number at any time, then calculates similarity Dynamic gene using relative entropy
(5b) is for user in the other term frequency-inverse document frequency score TF-IDF of each location categoryu,c, use corresponding phase
Like degree Dynamic geneCosine similarity calculation formula is adjusted, the similarity cos (u between two users is calculatedi,uj)；
(6) (3) are repeated the above steps extremely between the user and every other user that social network server recommends request
Step (5) obtains the similarity score of the user and other users, chooses the highest n user of score value and sends out as recommendation results
The user that request is recommended is given, n is the requested friend recommendation quantity of user that request is recommended.
The present invention has the following advantages that compared with prior art：
1) present invention recommends data available sparsity alleviating due to having used the semantic information in place of registering to classify
While, maintain higher recommendation precision.
2) present invention balances popular label due to having used term frequency-inverse document frequency model to quantify data of registering
To the preference in place and user itself.
3) of the invention since the distribution of behavior at any time of registering to user is analyzed, and by being signed between calculating user
The Dynamic gene for obtaining calculating user's similarity to the difference that behavior is distributed at any time takes full advantage of the geographical letter for information of registering
Breath and temporal information, to ensure that higher recommendation accuracy.
Detailed description of the invention
Fig. 1 is the communication system frame diagram that the present invention uses；
Fig. 2 is implementation flow chart of the invention；
Fig. 3 is the recommendation accuracy rate comparison diagram of suggested design used in the present invention with the suggested design for not considering time factor；
Fig. 4 is the recommendation recall rate comparison diagram of suggested design used in the present invention with the suggested design for not considering time factor.
Specific embodiment
Core of the invention thought is to be registered the language in place under location-based social networks scene by analyzing user
Adopted information classifies data of registering to alleviate the more sparse problem of data available, and the comprehensive analysis heat in place of registering
Degree, the preference of individual subscriber and user register the regularity of distribution of behavior at any time for carrying out friend recommendation to user, improve
The accuracy of recommendation.
Referring to Fig. 2, the present invention realizes that steps are as follows：
Step 1, communication system frame is established.
Referring to Fig.1, the communication system of this step foundation includes：User, location facilities, social network server, wherein using
Family and location facilities and social network server pass through mobile cellular net or WiFi carries out double-direction radio connection；
The user includes three application module, database module and GPS module functional modules；Application module is mainly used
In generating and sending information of registering to social network server；GPS module is mainly used for location facilities inquiring position information simultaneously
The geographical location information of user is provided to application module；Database module is mainly used for storage and management user and registers and good friend's letter
Breath；
The location facilities, mainly include GPS module, which is mainly used for ringing the position enquiring of user
It should and return to the geographical location information of user；
The social network server includes two functional modules of application module and database module；Application module is main
Analysis and arrangement is carried out for the information of registering to user and returns to recommendation results, database module for the request of the friend recommendation of user
It is mainly used for storing the semantic information of the register data and place of registering of user.
Step 2, information of registering is acquired.
User obtains the geographical location information where itself by location facilities, and by this geographical location information and at that time
Temporal information is sent to social network server and generates information of registering.
Step 3, social network server pre-processes the information of largely registering of storage.
(3a) is directed to the geographical location information in information of registering, and the position semantic database for searching server obtains and registers
Then the corresponding semantic information in place registers the information of registering of each user by the semantic information classification statistics in its place of registering
Number；
(3b) is directed to the temporal information registered in information, and the time is divided into 5 sections, and respectively 2：00-7：00,7：
00-11：00,11：00-17：00,17：00-22：00,22：00-2：00, counting user u each time zone on the classification c of place
Between on register number vis (c, u) and user the register number sum (c) total in the place classification, obtain user in place classification
Thus the frequency P (c, v) that registers=vis (c, v)/sum (c) of each time interval of c obtains user and registers in some place classification
The distribution of behavior at any time.
Step 4, social network server calculates each user u for the term frequency-inverse document frequency point of each place classification c
Value TF-IDFu,c。
(4a) social network server calculates user u and registers at the classification c of place the word frequency score value TF of informationu,c：
Wherein, | { u.vi:vi.D=c } | it is the number that user u registers at the classification c of place, | u.V | it is user u all
The total degree that place is registered；
(4b) social network server calculates the inverse document frequency score value IDF of each place classification cc：
Wherein, | U | indicate the quantity of all users in server, | { uj:c∈uj.D } | indicate that once classification c registered in place
User quantity；
Word frequency score value TF of (4c) social network server by user u at the classification c of placeu,cWith the inverse text of place classification c
Shelves frequency score IDFcIt is multiplied, obtains its term frequency-inverse document frequency score TF-IDF under the place classificationU, c=TFU, c×
IDFc。
Step 5, social network server calculates the similarity between user.
(5a) social network server calculates separately i-th of user uiWith j-th of user ujThe relative entropy D of classification c in place
(Pc||Qc) and j-th of user ujWith i-th of user uiRelative entropy D (the Q of classification c in placec||Pc), calculation formula is as follows：
Wherein, PcFor i-th of user uiThe probability distribution of classification c registered at any time, Q in placecFor j-th of user uj
The probability distribution of classification c registered at any time in place, i=1,2,3,4,5 respectively correspond step (3b) calculate user register with
5 time intervals divided when the probability distribution of time；
(5b) social network server calculates place classification c according to the result of (5a) and registers the Dynamic gene of information similarityIt is as follows：
The user that (5c) social network server is obtained in step (4c) is in the other term frequency-inverse document frequency of each location category
Score value TF-IDFu,cOn the basis of, cosine similarity calculation formula is adjusted using Dynamic gene, is calculated similar between two users
Spend cos (ui,uj)：
Wherein, IkFor i-th of user uiFor the term frequency-inverse document frequency score of place classification k, JkFor j-th of user uj
For the term frequency-inverse document frequency score of place classification k, FmaxFor the maximum value in all Dynamic genes.
Step 6, social network server makes final friend recommendation.
(3) are repeated the above steps between the user and every other user that social network server recommends request to step
(5), the similarity score of the user and other users is obtained, the highest n user of score value is chosen and is sent to as recommendation results
The user recommended is requested, n is the requested friend recommendation quantity of user that request is recommended.
Effect of the invention can be further illustrated by following emulation experiment：
1. experiment condition is arranged
Condition 1 obtains data set of registering in Foursquare, chooses the data composition data set of registering in wherein San Francisco
It closes, and chooses 50765 data of registering for the 100 most users of quantity that wherein register as the source data tested.
Condition 2, in Intel (R) Celeron (R) G540 processor, 7 Ultimate operating system of Windows it is desk-top
The result of test experiments on computer.
2. experiment content and result
Experiment 1：Filtered experimental data is divided into training set and test set, the present invention is run first on training set
Recommended program, 20 users most like with request recommended user will be obtained as the good friend of the user of request recommendation；With
Identical recommended program is run on test set afterwards, n most like user is taken to make recommendation respectively, n=1,2,3 ..., 20；
Then it is run on training set and test set and does not consider that the legacy buddy suggested design of time factor obtains corresponding buddy list
And recommendation results, finally the accuracy rate of two kinds of recommended methods is calculated, calculation formula is as follows：
Wherein R (u) indicates that, from the calculated buddy list of training set, T (u) indicates the friend recommendation made by test set
List, U indicate user all in data set.
Compare the two accuracy rate with the increased situation of change of commending friends quantity, as a result as shown in Figure 3.
It can be seen from figure 3 that after the good friend's quantity recommended is more than 5, it is of the invention used in friend recommendation scheme it is accurate
Rate is above the traditional friend recommendation scheme for not considering time factor.Under the usage scenario of reality, good friend's quantity of recommendation
Usually above 5, it can be said that the accuracy rate of friend recommendation scheme used in the present invention be substantially better than it is traditional when not considering
Between factor suggested design.
Experiment 2：Filtered experimental data is divided into training set and test set, the present invention is run first on training set
Recommended program, 20 users most like with request recommended user will be obtained as the good friend of the user of request recommendation；With
Identical recommended program is run on test set afterwards, n most like user is taken to make recommendation respectively, n=1,2,3 ..., 20；
Then it is run on training set and test set and does not consider that the legacy buddy suggested design of time factor obtains corresponding buddy list
And recommendation results, finally the recall rate of two kinds of recommended methods is calculated, calculation formula is as follows：
Wherein R (u) indicates that, from the calculated buddy list of training set, T (u) indicates the friend recommendation made by test set
List, U indicate user all in data set.
Compare the two recall rate with the increased situation of change of commending friends quantity, as a result as shown in Figure 4.
As seen from Figure 4, when good friend's negligible amounts of recommendation, friend recommendation scheme used in the present invention is not examined with traditional
The recall rate difference for considering the friend recommendation scheme of time factor is little, with the increase of commending friends quantity, used in the present invention
Scheme recall rate is more than traditional friend recommendation scheme for not considering the time gradually.Also due under the usage scenario of reality,
Good friend's quantity of recommendation is usually more, it can be said that the recall rate of friend recommendation scheme used in the present invention is substantially better than tradition
The suggested design for not considering time factor.
In conclusion the present invention is superior to traditional friend recommendation system in terms of the accuracy rate of recommendation results and recall rate
System, can provide quality higher friend recommendation service for the user based on position social networks.
Claims (6)
1. a kind of friend recommendation method in the social networks of position based on position and time, including：
(1) the communication system frame being made of user, location facilities, social network server is established, wherein：
User, for being communicated by mobile cellular network or WiFi with location facilities and social network server；
Location facilities realize positioning for collaborative user's GPS module；
Social network server carries out friend recommendation to user for storing the location information of user, and on this basis；
(2) longitude and latitude of position and current temporal information are sent to social network server by user, generate information of registering；
(3) social network server pre-processes the information of registering of storage：
(3a) is directed to the longitude and latitude in information of registering, and the position semantic database acquisition for searching server is corresponding with place of registering
Then the information of registering of each user is counted number of registering by the semantic information classification in its place of registering by semantic information；
The temporal information that (3b) is directed in information of registering counts each user in each location category on the basis of step (3a)
The distribution of behavior at any time of registering under not；
(4) according to user in the number of registering of each place classification, each user u of social network server calculating is for each place
The term frequency-inverse document frequency score TF-IDF of classification cu,c；
The number of registering of (4a) according to user u under each place classification, social network server calculate user u in each location category
The word frequency score value TF for information of registering under other cu,c；
The number of registering of (4b) according to all users under each place classification, social network server calculate each place classification c
Inverse document frequency score value IDFc；
Word frequency score value TF of (4c) social network server by user u at the classification c of placeu,cWith the inverse document frequency of place classification c
Rate score value IDFcIt is multiplied, obtains its term frequency-inverse document frequency score TF-IDF under the place classificationu,c=TFu,c×IDFc；
(5) social network server calculates the similarity between user：
(5a) is registered the distribution of behavior at any time using user under each place classification, calculates two users in identical location category
Register the relative entropy D (P | | Q) of distribution under other, and wherein P and Q is two user uiAnd ujIdentical place classification register number with
Then the probability distribution of time calculates similarity Dynamic gene using relative entropy
(5b) is for user in the other term frequency-inverse document frequency score TF-IDF of each location categoryu,c, use corresponding similarity
Dynamic geneCosine similarity calculation formula is adjusted, the similarity cos (u between two users is calculatedi,uj)；
(6) (3) are repeated the above steps to step between the user and every other user that social network server recommends request
(5), the similarity score of the user and other users is obtained, the highest n user of score value is chosen and is sent to as recommendation results
The user recommended is requested, n is the requested friend recommendation quantity of user that request is recommended.
2. being by one according to the method described in claim 1, wherein step (3b) counting user registers behavior when being distributed at any time
It is divided into 5 sections to the time, and respectively 2：00-7：00,7：00-11：00,11：00-17：00,17：00-22：00,22：00-
2：Register number vis (c, u) and the user of 00, counting user u on the classification c of place on each time interval are in the place classification
Total number sum (c) that registers, obtain user each time interval of place classification c the frequency P (c, v) that registers=vis (c, v)/
Sum (c) obtains user and registers the distribution of behavior at any time in some place classification.
3. according to the method described in claim 1, wherein social network server calculates user u in each place in step (4a)
The word frequency score value TF for information of registering under classification cu,c, calculated by following formula：
Wherein, | { u.vi:vi.D=c } | it is the number that user u registers at the classification c of place, | u.V | it is user u in all places
The total degree registered.
4. according to the method described in claim 1, the wherein inverse document frequency score value IDF of step (4b) place classification ccUnder
Formula calculates：
Wherein, | U | indicate the quantity of all users in server, | { uj:c∈uj.D } | indicate the use that once classification c registers in place
The quantity at family.
5. according to the method described in claim 2, wherein calculating two user u in step (5a)iAnd ujPlace classification c is signed
Dynamic gene when to information similarityIt carries out as follows：
(5a1) calculates separately i-th of user uiWith j-th of user ujRelative entropy D (the P of classification c in placec||Qc) used with j-th
Family ujWith i-th of user uiRelative entropy D (the Q of classification c in placec||Pc)：
Wherein, PcFor i-th of user uiThe probability distribution of classification c registered at any time, Q in placecFor j-th of user ujIn place
The probability distribution of classification c registered at any time, i=1,2,3,4,5, which respectively correspond step (3b) calculating user, registers at any time
5 time intervals divided when probability distribution；
(5a2) calculates place classification c according to the result of (5a1) and registers the Dynamic gene of information similarityIt is as follows：
6. according to the method described in claim 1, wherein step (5b) utilizes similarity between cosine similarity two users of calculating
cos(ui,uj), it is calculate by the following formula：
Wherein, IkFor i-th of user uiFor the term frequency-inverse document frequency score of place classification k, JkFor j-th of user ujFor
The term frequency-inverse document frequency score of place classification k, FmaxFor the maximum value in all Dynamic genes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610051508.3A CN105791085B (en) | 2016-01-26 | 2016-01-26 | Friend recommendation method in the social networks of position based on position and time |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610051508.3A CN105791085B (en) | 2016-01-26 | 2016-01-26 | Friend recommendation method in the social networks of position based on position and time |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105791085A CN105791085A (en) | 2016-07-20 |
| CN105791085B true CN105791085B (en) | 2018-11-20 |
Family
ID=56403089
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610051508.3A Active CN105791085B (en) | 2016-01-26 | 2016-01-26 | Friend recommendation method in the social networks of position based on position and time |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105791085B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106373016A (en) * | 2016-09-26 | 2017-02-01 | 珠海市魅族科技有限公司 | Method and apparatus for establishing online social relations |
| CN106570764A (en) * | 2016-11-09 | 2017-04-19 | 广州杰赛科技股份有限公司 | User relationship predicting method and device |
| CN106910135A (en) * | 2017-01-25 | 2017-06-30 | 百度在线网络技术（北京）有限公司 | User recommends method and device |
| CN108574715A (en) * | 2017-03-14 | 2018-09-25 | 广州市动景计算机科技有限公司 | Information recommendation method, apparatus and system |
| CN107181672A (en) * | 2017-06-09 | 2017-09-19 | 西安电子科技大学 | The friend recommendation method based on Annual distribution relative entropy in the social networks of position |
| CN107657015B (en) * | 2017-09-26 | 2021-03-19 | 北京邮电大学 | Interest point recommendation method and device, electronic equipment and storage medium |
| CN107766553A (en) * | 2017-11-02 | 2018-03-06 | 成都金川田农机制造有限公司 | Based on text mining by weight colony portrait method |
| CN107995099A (en) * | 2017-11-24 | 2018-05-04 | 广东欧珀移动通信有限公司 | Friend recommendation method and device |
| CN110059795A (en) * | 2018-01-18 | 2019-07-26 | 中国科学院声学研究所 | A kind of mobile subscriber's node networking method merging geographical location and temporal characteristics |
| CN109035050A (en) * | 2018-07-25 | 2018-12-18 | 安徽新华学院 | A kind of location-based social recommendation system |
| CN110049447B (en) * | 2019-04-12 | 2021-02-05 | 桂林电子科技大学 | Position information-based partnership analysis method |
| CN114579879B (en) * | 2022-05-06 | 2022-07-26 | 南方科技大学 | Friend recommendation method, device, equipment and storage medium |
| CN117131240A (en) * | 2023-02-10 | 2023-11-28 | 荣耀终端有限公司 | Service recommendation method, electronic device and computer readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102594905A (en) * | 2012-03-07 | 2012-07-18 | 南京邮电大学 | Method for recommending social network position interest points based on scene |
| CN102752708A (en) * | 2011-04-20 | 2012-10-24 | 曹晓刚 | Parallel friend recommendation system and method capable of serving on basis of geographic positions |
| CN104881459A (en) * | 2015-05-22 | 2015-09-02 | 电子科技大学 | Friend recommendation method of mobile social network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120109752A1 (en) * | 2009-08-19 | 2012-05-03 | Vitrue, Inc. | Systems and methods for delivering targeted content to a consumer's mobile device based on the consumer's physical location and social media memberships |
-
2016
- 2016-01-26 CN CN201610051508.3A patent/CN105791085B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102752708A (en) * | 2011-04-20 | 2012-10-24 | 曹晓刚 | Parallel friend recommendation system and method capable of serving on basis of geographic positions |
| CN102594905A (en) * | 2012-03-07 | 2012-07-18 | 南京邮电大学 | Method for recommending social network position interest points based on scene |
| CN104881459A (en) * | 2015-05-22 | 2015-09-02 | 电子科技大学 | Friend recommendation method of mobile social network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105791085A (en) | 2016-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105791085B (en) | Friend recommendation method in the social networks of position based on position and time | |
| US9330174B1 (en) | Determining topics of interest | |
| Parra-López et al. | Travellers 2.0: Motivation, opportunity and ability to use social media | |
| Liu et al. | Point-of-interest recommendation in location based social networks with topic and location awareness | |
| US8667009B2 (en) | Systems and methods for utilizing and searching social network information | |
| US9208444B1 (en) | Determining happiness indices | |
| Leung et al. | Attracting user social media engagement: A study of three budget airlines Facebook pages | |
| US20090106307A1 (en) | System of a knowledge management and networking environment and method for providing advanced functions therefor | |
| CN108550068A (en) | A kind of individual commodity recommendation method and system based on user behavior analysis | |
| US20100004975A1 (en) | System and method for leveraging proximity data in a web-based socially-enabled knowledge networking environment | |
| US9262481B1 (en) | Determining popular venues | |
| Li et al. | A multi-dimensional context-aware recommendation approach based on improved random forest algorithm | |
| US20130006881A1 (en) | Method of identifying relevant user feedback | |
| Pai et al. | Understanding tourist preferences for travel packages: a conjoint analysis approach | |
| Smith | Metrics, locations, and lift: Mobile location analytics and the production of second-order geodemographics | |
| Wang et al. | Group-based personalized location recommendation on social networks | |
| Liu et al. | Personalised context-aware re-ranking in recommender system | |
| CN105978729B (en) | A kind of cellphone information supplying system and method based on user's internet log and position | |
| CN107181672A (en) | The friend recommendation method based on Annual distribution relative entropy in the social networks of position | |
| Yu | Cloud storage–based personalized sports activity management in Internet plus O2O sports community | |
| CN105208033B (en) | A kind of colony's auxiliary based on intelligent terminal scene recommends method and system | |
| Gao et al. | Personalized recommendation method of POI based on deep neural network | |
| Li et al. | A personalized recommendation algorithm for college books based on user interest | |
| Yu et al. | Personalized POI groups recommendation in location-based social networks | |
| Yoon et al. | An empirical comparison of vignette and virtual reality experiments in tourism research |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |