CN106971345A - A kind of location recommendation method based on position social networks - Google Patents
A kind of location recommendation method based on position social networks Download PDFInfo
- Publication number
- CN106971345A CN106971345A CN201610012330.1A CN201610012330A CN106971345A CN 106971345 A CN106971345 A CN 106971345A CN 201610012330 A CN201610012330 A CN 201610012330A CN 106971345 A CN106971345 A CN 106971345A
- Authority
- CN
- China
- Prior art keywords
- user
- place
- active user
- friend
- similarity
- 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
Classifications
-
- 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 invention discloses a kind of location recommendation method based on position social networks, comprise the following steps：Set up location-based social networks LBSNs；Calculate active user u1With friend u2Similarity and u on record of registering1And u2Similarity on social networks；Consider above two similarity, calculating obtains u1With u2Similarity；Obtain K and u1Similarity highest friend is used as nearest-neighbors；Accessed and u for nearest-neighbors1The place p not accessed, calculates u1P frequency may be accessed；The multiple places of selecting frequency highest are place Candidate Set；Calculate active user u1Candidate locations p possibility is accessed, the maximum multiple candidate locations of possibility are selected as recommendation place.It is openness and produce the recommendation results of better quality that this method can make up matrix in the case where user items rating matrix is sparse.
Description
Technical field
The present invention relates to commending system and based on position field of social network.
Background technology
Recently, the development (GPS, Wi-Fi) of mobile device technology makes geographical location information and traditional online social networks
Be combined into possibility, traditional online social networks enters to turn to location-based social networks (LBSNs).Geographical location information
The physical dimension of network shortage can be served as.In this new field, user can use base whenever and wherever possible by mobile device
Service in position.In addition, user can share the information related to place with other people.
In actual application, the rating matrix of user-project is generally all, than sparse, therefore, much to recommend knot
It is really unsatisfactory.
The content of the invention
In view of this, the invention provides a kind of location recommendation method based on position social networks, this method can be
Matrix can be made up in the case that user-project rating matrix is sparse openness and produce the recommendation results of better quality.
In order to achieve the above object, technical scheme comprises the following steps：
Step 1, the user's set set up in location-based social networks LBSNs, the ground point set of user, the label of user
To set and friend's set of each user.
Step 2, for active user u1, obtain its register integrate be combined into C, point set as L, friend collection be combined into F；
Being registered from it, it is equal that each data in set a LN, LN for describing user-place-number of times of registering is obtained in collecting C
Including active user u1ID, active user u1Place ID and active user u1ID registered in each place ID history
Number of times.
Every number in the set FN, FN of a description user-user-common friends quantity is obtained from friend's set F
According to including active user u1And its ID, the active user u of friend1The quantity of common friends between its friend.
Step 3, active user u1With friend u2Similarity on record of registeringFor：
Wherein, fU, lRepresent register number of times of the user u in place l；Wherein u2∈F。
Step 4, active user u1With friend u2Similarity on social networksFor：
U represents the set of user,Represent user u1With the quantity of user's u common friends.
Step 5, the result according to step 3 and step 4, calculating obtain active user u1With friend u2Similarity be
Wherein,Represent active user u1With friend u2Geographical position between distance.
Step 6, the result of calculation according to step 5, obtain K and active user u1Similarity highest friend is as nearest
Neighbours.
Step 7, pass through active user u1K nearest-neighbors, calculating obtain place set P, set P is by active user u1
Nearest-neighbors accessed and u1The place not accessed is constituted；For each place p in set P, current use is calculated
Family u1Possible access locations p frequency For active user u1With accessed place p most
Neighbour user upSimilarity,For upIn place, p history is registered number of times.
Choose preceding 50 pre- measured frequency highest places and be used as place Candidate Set Lc。
Step 8, traversal LcIn each candidate locations lc, calculate active user u1Access candidate locations lcPossibility,
The maximum multiple candidate locations of possibility are selected as recommendation place.
The set L in the place that existing subscriber the accessed and geographical position distance set D in each pair place in set.Calculate lc
With each place l in ground point set LiDi, i ∈ [1, n], n is the number in place in set L.
Then active user u1Access candidate locations lcProbability be
D is active user u1The geographical distance sample set of access；The geographical position distance of place between any two in set L
Set, h is bandwidth, according to following bandwidth equations：
For active user u1The geographical distance sample standard deviation of access.
Active user u1Candidate locations l may be accessedcProbability be
Then active user u1Access candidate locations lcPossibility be：P (u, l)=pCF(u, l) × pdistribution(u, l).
Further, active user u1It is not cold start-up user, then the D is the geography of place between any two in set L
Positional distance set；If active user u1For cold start-up user, then the D includes u1Between each place in set L and
The geographical position distance of place between any two in set L.
Beneficial effect：
1st, method provided by the present invention is to be recorded and social networks using Similarity-Weighted algorithm to being registered based on history
Similarity be weighted.Under normal circumstances, the probability in user's access identical place of friend is higher each other.Distance between house
The probability that two nearer users access identical place is higher.So, fusion based on user register record calculate it is similar
During degree and the similarity calculated based on user social contact relation, the present invention is according to the distance between two user's houses, certainly
Two kinds of similarities of dynamic adjustment proportion shared in proposed algorithm.
2nd, the proposed algorithm that Density Estimator and Collaborative Filtering Recommendation Algorithm are combined by the present invention.It is existing to be based on geography
The calculation that is influenceed on geographic factor of algorithm of position influence can not Coping with Reality situation well.The present invention recommends for user
During place, personalisation process is carried out to the geographic influence of the behavior of registering of each user.The present invention calculates user's history label first
To the distance of intersite, then choose non-parametric test method-Density Estimator prediction history and register the distribution of intersite distance
Function, finally, calculates the probability for accessing candidate locations.
3rd, the present invention using based on mixing collaborative filtering recommendation method, solve in traditional collaborative filtering due to user-
Project rating matrix is relatively sparse, cold start-up problem, and the recommendation results caused are inaccurate, recommend the problem of precision is not high；Separately
On the one hand, a set of practicable system schema is formd, for qualified input, can produce and preferably push away relatively
Recommend result.
Brief description of the drawings
Fig. 1 for the present invention in recommendation method SKHCF algorithms flow.
Embodiment
The present invention will now be described in detail with reference to the accompanying drawings and examples.
A kind of location recommendation method based on position social networks is present embodiments provided, is comprised the following steps：
Step 1, the user's set set up in location-based social networks LBSNs, the ground point set of user, the label of user
To set and friend's set of each user.
For active user u1, its register integrate be combined into C, point set as L, friend collection be combined into F；
Being registered from it, it is equal that each data in set a LN, LN for describing user-place-number of times of registering is obtained in collecting C
Including active user u1ID, active user u1Place ID and active user u1ID registered in each place ID history
Number of times.
Every number in the set FN, FN of a description user-user-common friends quantity is obtained from friend's set F
According to including active user u1And its ID, the active user u of friend1The quantity of common friends between its friend.
For step 2, the relative user registered in different location, the user's similarity registered in identical place is higher.In addition,
The frequency registered in identical place is more similar, and similarity is got over, therefore the present embodiment considers the label of user during similarity is calculated
To frequency.Active user u1With friend u2Similarity on record of registeringFor：
Wherein, fU, lRepresent register number of times of the user u in place l；Wherein u2∈F。
Similarity between step 3, friend is higher than the similarity between non-friend User, common friends number
The similarity of two higher users is higher.Active user u1With friend u2Similarity on social networksFor：U represents the set of user,Represent user u1With user's u common friends
Quantity.
Step 4, the result according to step 2 and step 3, calculating obtain active user u1With friend u2Similarity be
Wherein,Represent active user u1With friend u2Geographical position between distance.
House distance between i.e. two friends it is nearer, the probability for accessing identical place is higher.Conversely, when house distance is big
The probability that identical place is accessed after to a certain extent, between friend is approximately zero.
Step 5, the result of calculation according to step 4, obtain K and active user u1Similarity highest friend is as nearest
Neighbours.
Step 6, pass through active user u1K nearest-neighbors, calculating obtain place set P, set P is by active user u1
Nearest-neighbors accessed and u1The place not accessed is constituted；For each place p in set P, current use is calculated
Family u1Possible access locations p frequency For active user u1With accessed place p most
Neighbour user upSimilarity,For upIn place, p history is registered number of times.
Multiple pre- measured frequency highest places are used as place Candidate Set L before choosingc, the present embodiment have chosen 50.
Step 7, traversal LcIn each candidate locations lc, calculate active user u1Access candidate locations lcPossibility,
The maximum multiple candidate locations of possibility are selected as recommendation place.
Calculate lcWith each place l in ground point set LiDi, i ∈ [1, n], n is the number in place in set L.
Then active user u1Access candidate locations lcProbability be
D is active user u1The geographical distance sample set of access；The geographical position distance of place between any two in set L
Set, h is bandwidth.
Active user u1Candidate locations l may be accessedcProbability be
Then active user u1Access candidate locations lcPossibility be：P (u, l)=pCF(u, l) × pdistribution(u, l).
Further, active user u1It is not cold start-up user, then the D is the geography of place between any two in set L
Positional distance set；If active user u1For cold start-up user, then the D includes u1Between each place in set L and
The geographical position distance of place between any two in set L.
To sum up, presently preferred embodiments of the present invention is these are only, is not intended to limit the scope of the present invention.It is all
Within the spirit and principles in the present invention, any modification, equivalent substitution and improvements made etc. should be included in the protection of the present invention
Within the scope of.
Claims (3)
1. a kind of location recommendation method based on position social networks, it is characterised in that comprise the following steps：
Step 1, set up in location-based social networks LBSNs, wherein LBSNs user and gather, the ground point set of user, user
Register set and friend's set of each user；For active user u1, its register integrate be combined into C, point set as L, friend
Friend's collection is combined into F；
Step 2, active user u1With friend u2Similarity on record of registeringFor：
Wherein, fU, lRepresent register number of times of the user u in place l；Wherein u2∈F；
Step 3, active user u1With friend u2Similarity on social networksFor：
U represents the set of user,Represent user u1With user u2The quantity of common friends；
Step 4, the result according to step 2 and step 3, calculating obtain active user u1With friend u2Similarity be
Wherein,Represent active user u1With friend u2Geographical position between distance；
Step 5, the result of calculation according to step 4, obtain K and active user u1Similarity highest friend is used as nearest-neighbors；
Step 6, pass through active user u1K nearest-neighbors, calculating obtain place set P, set P is by active user u1Most
Neighbour occupies and accessed and u1The place not accessed is constituted；For each place p in set P, active user u is calculated1
Possible access locations p frequency For active user u1With the arest neighbors for accessing place p
User upSimilarity,For upIn place, p history is registered number of times；
Multiple frequency highest places are used as place Candidate Set L before choosingc；
Step 7, traversal lcIn each candidate locations lc, calculate active user u1Access candidate locations lcPossibility, select
The maximum multiple candidate locations of possibility are used as recommendation place；
Calculate lcWith each place l in ground point set LiDi, i ∈ [1, n], n is the number in place in set L；
Then active user u1Access candidate locations lcProbability be
D is active user u1The geographical distance sample set of access；The geographical position distance set of place between any two in set L,
H is bandwidth, according to following bandwidth equations：
For active user u1The geographical distance sample standard deviation of access；
Active user u1Candidate locations l may be accessedcProbability be
Then active user u1Access candidate locations lcPossibility be：P (u, 1)=pCF(u, 1) × pdistribution(u, l).
2. a kind of location recommendation method based on position social networks as claimed in claim 1, it is characterised in that registered from it
Each data in the set LN, LN of description user-place-number of times of registering is obtained in collection C includes active user u1's
ID, active user u1Place ID and active user u1ID registered number of times in each place ID history；
The every data obtained from friend's set F in the set FN, FN of a description user-user-common friends quantity is equal
Including active user u1And its ID, the active user u of friend1The quantity of common friends between its friend.
3. a kind of location recommendation method based on position social networks as claimed in claim 1, it is characterised in that active user
u1It is not cold start-up user, then the D is the geographical position distance set of place between any two in set L；If active user u1
For cold start-up user, then the D includes u1The geography of place between any two between each place in set L and in set L
Positional distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610012330.1A CN106971345A (en) | 2016-01-08 | 2016-01-08 | A kind of location recommendation method based on position social networks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610012330.1A CN106971345A (en) | 2016-01-08 | 2016-01-08 | A kind of location recommendation method based on position social networks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106971345A true CN106971345A (en) | 2017-07-21 |
Family
ID=59334167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610012330.1A Pending CN106971345A (en) | 2016-01-08 | 2016-01-08 | A kind of location recommendation method based on position social networks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106971345A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107766462A (en) * | 2017-09-28 | 2018-03-06 | 重庆大学 | Point of interest based on user preference, social credit worthiness and geographical position recommends method |
| CN108600961A (en) * | 2018-03-23 | 2018-09-28 | 广州杰赛科技股份有限公司 | Preparation method and device, equipment, the storage medium of user's similarity |
| CN109902883A (en) * | 2019-03-25 | 2019-06-18 | 重庆邮电大学 | A kind of position predicting method of registering based on personalized level Density Estimator |
| CN110020225A (en) * | 2017-09-06 | 2019-07-16 | 丰田自动车株式会社 | Information processing unit, information processing system and information processing method |
| CN110968771A (en) * | 2018-09-29 | 2020-04-07 | 北京淘友天下技术有限公司 | Position recommendation cold start method and system based on friendship |
| CN112508725A (en) * | 2020-12-04 | 2021-03-16 | 首都师范大学 | Community structure-based location awareness influence maximization method |
| CN114417174A (en) * | 2022-03-23 | 2022-04-29 | 腾讯科技（深圳）有限公司 | Content recommendation method, device, equipment and computer storage medium |
| CN114637912A (en) * | 2022-03-08 | 2022-06-17 | 浙江工商大学 | Friend recommendation method and device based on high-coverage community discovery in location social network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100153292A1 (en) * | 2008-12-11 | 2010-06-17 | Microsoft Corporation | Making Friend and Location Recommendations Based on Location Similarities |
| 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 |
| US20120233238A1 (en) * | 2011-03-07 | 2012-09-13 | David Edward Braginsky | Dynamic Recommendations in Geo-Social Networking System |
| CN104317900A (en) * | 2014-10-24 | 2015-01-28 | 重庆邮电大学 | Multiattribute collaborative filtering recommendation method oriented to social network |
| CN105224545A (en) * | 2014-06-03 | 2016-01-06 | 华为技术有限公司 | A kind of position recommend method and device |
-
2016
- 2016-01-08 CN CN201610012330.1A patent/CN106971345A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100153292A1 (en) * | 2008-12-11 | 2010-06-17 | Microsoft Corporation | Making Friend and Location Recommendations Based on Location Similarities |
| 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 |
| US20120233238A1 (en) * | 2011-03-07 | 2012-09-13 | David Edward Braginsky | Dynamic Recommendations in Geo-Social Networking System |
| CN105224545A (en) * | 2014-06-03 | 2016-01-06 | 华为技术有限公司 | A kind of position recommend method and device |
| CN104317900A (en) * | 2014-10-24 | 2015-01-28 | 重庆邮电大学 | Multiattribute collaborative filtering recommendation method oriented to social network |
Non-Patent Citations (2)
| Title |
|---|
| 周而重等: "一种基于用户网络签到行为的地点推荐方法", 《计算机科学》 * |
| 袁适之等: "一种基于位置社交网络的地点推荐算法", 《计算机应用研究》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110020225A (en) * | 2017-09-06 | 2019-07-16 | 丰田自动车株式会社 | Information processing unit, information processing system and information processing method |
| CN107766462A (en) * | 2017-09-28 | 2018-03-06 | 重庆大学 | Point of interest based on user preference, social credit worthiness and geographical position recommends method |
| CN107766462B (en) * | 2017-09-28 | 2020-03-13 | 重庆大学 | Interest point recommendation method based on user preference, social reputation and geographic position |
| CN108600961A (en) * | 2018-03-23 | 2018-09-28 | 广州杰赛科技股份有限公司 | Preparation method and device, equipment, the storage medium of user's similarity |
| CN110968771A (en) * | 2018-09-29 | 2020-04-07 | 北京淘友天下技术有限公司 | Position recommendation cold start method and system based on friendship |
| CN109902883A (en) * | 2019-03-25 | 2019-06-18 | 重庆邮电大学 | A kind of position predicting method of registering based on personalized level Density Estimator |
| CN112508725A (en) * | 2020-12-04 | 2021-03-16 | 首都师范大学 | Community structure-based location awareness influence maximization method |
| CN112508725B (en) * | 2020-12-04 | 2023-02-17 | 首都师范大学 | Community structure-based location awareness influence maximization method |
| CN114637912A (en) * | 2022-03-08 | 2022-06-17 | 浙江工商大学 | Friend recommendation method and device based on high-coverage community discovery in location social network |
| CN114637912B (en) * | 2022-03-08 | 2024-05-03 | 浙江工商大学 | Friend recommendation method and device based on high-coverage community discovery in location social network |
| CN114417174A (en) * | 2022-03-23 | 2022-04-29 | 腾讯科技（深圳）有限公司 | Content recommendation method, device, equipment and computer storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106971345A (en) | A kind of location recommendation method based on position social networks | |
| Bao et al. | Location-based and preference-aware recommendation using sparse geo-social networking data | |
| CN103617547B (en) | A kind of business recommended method and system | |
| US8566309B2 (en) | Monitoring hashtags in micro-blog posts to provide one or more crowd-based features | |
| US20160034712A1 (en) | System and method for event-related content discovery, curation, and presentation | |
| Quercia et al. | Friendsensing: recommending friends using mobile phones | |
| US8789133B2 (en) | Location based content filtering and dynamic policy | |
| CN105791085B (en) | Friend recommendation method in the social networks of position based on position and time | |
| US8843463B2 (en) | Providing content by using a social network | |
| CN103905471B (en) | Information-pushing method, server and the social networks of social networks | |
| US20130282723A1 (en) | Maintaining A Historical Record Of Anonymized User Profile Data By Location For Users In A Mobile Environment | |
| US8688735B2 (en) | Determining potential relationships between individuals | |
| KR20140071418A (en) | Providing mobile advertising using data networks based on intelligence data associated with internet-connected devices | |
| Li et al. | Multi-layered friendship modeling for location-based mobile social networks | |
| CN104391853A (en) | POI (point of interest) recommending method, POI information processing method and server | |
| CN105163142B (en) | A kind of user preference determines method, video recommendation method and system | |
| US20130311283A1 (en) | Data mining method for social network of terminal user and related methods, apparatuses and systems | |
| CN108804551A (en) | It is a kind of to take into account diversity and personalized space point of interest recommendation method | |
| CN107341261A (en) | A kind of point of interest of facing position social networks recommends method | |
| CN107038649B (en) | Friend recommendation method and device for terminal user | |
| CN102316167A (en) | Website recommending method, system thereof and network server | |
| CN110322067A (en) | Location of mobile users prediction technique based on factor graph model | |
| CN105978729B (en) | A kind of cellphone information supplying system and method based on user's internet log and position | |
| CN109525933B (en) | Location privacy protection method and device | |
| CN105956061B (en) | Method and device for determining similarity between users |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170721 |
|
| WD01 | Invention patent application deemed withdrawn after publication |