CN112948711A - Social network friend recommendation system based on geographical position sharing track - Google Patents
Social network friend recommendation system based on geographical position sharing track Download PDFInfo
- Publication number
- CN112948711A CN112948711A CN202110314101.6A CN202110314101A CN112948711A CN 112948711 A CN112948711 A CN 112948711A CN 202110314101 A CN202110314101 A CN 202110314101A CN 112948711 A CN112948711 A CN 112948711A
- Authority
- CN
- China
- Prior art keywords
- user
- users
- sharing
- friend
- geographical position
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a social network friend recommendation system based on a geographical position sharing track, which belongs to the field of friend recommendation systems and mainly comprises a data collection system, a user characteristic representation system, a user matching degree calculation system and a friend recommendation system, wherein in the step of the user characteristic representation system, the front and back sequence of the user geographical position sharing, namely the geographical position sharing track, is integrated into the user characteristic representation system, so that the interests, hobbies and behavior habits of users are more fully reflected.
Description
Technical Field
The invention relates to the field of friend recommendation systems, in particular to a social network friend recommendation system based on a geographical position sharing track.
Background
With the rise and development of mobile devices, geolocation and social media, social network users can share geographical location information and related activities online to friends on social software, and software providing such geographical location sharing functions includes WeChat, strange, FourSquare and Yelp, etc. This functionality links the virtual cyberspace with the physical geospatial, facilitating social relationships and communication dynamics between users via the users' geospatial information. On the basis, the friend recommendation system can recommend users with similar interests to each other, so that the formation of a social network is accelerated, the rapid development of an online community is promoted, the frequency of using social software by the users is improved, and the friend recommendation system plays a vital role in improving the user stickiness and developing the number of the users.
In the prior art, rafailis and mastic (rafailis, Dimitrios, and facial creatino), "Friend reception in location-based Social network view deep pair learning," 2018IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM).
In the prior art, patent application No. 201510425193.X provides a method, a device and a system for recommending a terminal based on a movement track, and the patent uses a specific device to continuously acquire the movement track of a user and recommends friends to the user based on the contact ratio of the movement track in the same time.
The scheme depends on a specific terminal device, and the application range of the method is limited; the scheme needs to continuously acquire the moving track of the user, the privacy requirement of the user is difficult to guarantee, and if the user temporarily stops sharing the moving track midway, the scheme cannot extract the user characteristics and realize a friend recommendation system; in practical situations, two users visiting the same place at different times can reflect that the two users have common interests, but the method only considers whether the two users can simultaneously travel a section of path when recommending friends, which results in low precision of the friend recommendation system of the method.
Disclosure of Invention
The embodiment of the invention provides a social network friend recommendation system based on a geographical position sharing track, and aims to solve the technical problems that a friend recommendation system in the prior art is low in recommendation success rate, mismatched in users and not beneficial to geographical position sharing.
The embodiment of the invention adopts the following technical scheme: a social network friend recommendation system based on a geographical position sharing track comprises a data collection system, a user characteristic representation system, a user matching degree calculation system and a friend recommendation system, wherein user information is collected by the data collection system, then the user information is distinguished and represented according to the user characteristic representation system, simulation calculation is carried out according to the user matching degree, and the friend recommendation system is provided for a user according to a calculation result.
Furthermore, the data collection system collects geographical location sharing information of the social users and social information crawling source data through social software, the geographical location sharing information comprises sharing time, geographical location longitude and latitude, place merchant names, place merchant types and the like, and the social information comprises friend lists of the users and friend adding time.
Furthermore, the user characteristic representation system adopts an unsupervised learning method, each user and each place are respectively represented as a vector according to the characteristics of the user geographical position sharing track, the sharing time and the like, and the theoretical basis of the model is that the geographical position shared by the users can pass through the user sharing habit and the places before and after the geographical position.
Furthermore, a deep neural network model is arranged in the user matching degree calculation system, the matching degree of each pair of users is calculated through the deep neural network model, the model is input by two users, and the embedded vector representations of the two users are respectively calculated through a user embedded matrix U; then splicing the two vectors, and calling the spliced vector as H0On top of which a q-layer fully connected neural network is connectedH1，……，Hq，HqI.e. an implicit expression of the relationship between two users, in the q-layer fully-connected neural network, a ReLU function is used as an activation function of the fully-connected neural networks, namely Hk＝ReLU(WkHk-1+bk) (ii) a And finally, calculating the friend matching degrees of the two users by using a Sigmoid function, namely the friend matching degree Fij＝sigmoid(Wq+1Hq+bq+1)。
Furthermore, the friend recommending system recommends m friends for each user, and for one user u, sorts the users who are not friends of the user according to the matching degree, selects the top m users from the sorted users, and recommends the users to the user u.
The embodiment of the invention adopts at least one technical scheme which can achieve the following beneficial effects:
the system mainly comprises a data collection system, a user characteristic representation system, a user matching degree calculation system and a friend recommendation system, wherein in the step of the user characteristic representation system, the front-back sequence of user geographical position sharing, namely a geographical position sharing track, is integrated into the user characteristic representation system, and the interests, hobbies and behavior habits of users are more fully reflected.
Secondly, the user can enjoy the social recommendation function provided by the scheme by using a common smart phone without depending on a specific positioning or track tracking device. Meanwhile, the method ensures the privacy of the user, and the user can share the geographical position only when the user wants to share the geographical position. Even if the user track is discontinuous, the method can also excavate the interests, hobbies and behavior habits of the user through long-term geographical position sharing of the user, and the effective friend recommendation system of the user is realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a diagram of a deep learning model of a user feature representation system in accordance with the present invention;
FIG. 3 is a diagram of a deep neural network model.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical solutions provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
A social network friend recommendation system based on a geographical position sharing track comprises a data collection system, a user characteristic representation system, a user matching degree calculation system and a friend recommendation system, wherein user information is collected by the data collection system, then the user information is distinguished and represented according to the user characteristic representation system, simulation calculation is carried out according to the user matching degree, and the friend recommendation system is provided for a user according to a calculation result.
The method does not depend on a specific positioning or track tracking device, and a user can enjoy the social recommendation function provided by the scheme by using a common smart phone. Meanwhile, the method ensures the privacy of the user, and the user can share the geographical position only when the user wants to share the geographical position. Even if the user track is discontinuous, the method can also excavate the interests, hobbies and behavior habits of the user through long-term geographical position sharing of the user, and the effective friend recommendation system of the user is realized.
A data collection system: the method comprises the steps of collecting geographical location sharing information and social information of social users through crawling source data or through social software, wherein the geographical location sharing information comprises sharing time, geographical location longitude and latitude, place merchant names, place merchant types and the like, and the social information comprises a friend list of each user and time of adding friends.
User feature representation system: the method adopts an unsupervised learning method, and respectively represents each user and each place into a vector according to the characteristics of the user geographical position sharing track, the sharing time and the like, and the theoretical basis of the model is that the geographical position shared by the user can be predicted through the user sharing habit and the places before and after the geographical position;
for example, if a geographical location sharing sequence of users of an office group is (mcdonald, station,.
For a given site, the sites that define its back and forth visits are:
C(u，li)＝(li-j，...，li-1，li+1，...，li+j)
wherein j is the window size, and the geographic information sharing track T of a given user uuThe feature represents that the learning model aims to give a time of geographic information sharing and a place of front and back visit, and the computing user u calculates the place l of the user u at the timeiProbability of sharing geographic information, i.e. P (l)i|u，ti，C(u，li) Or P (given location | user, time, location visited before and after). Because the geographic position sharing data set does not contain negative examples, the negative sampling technology is adopted to randomly slave IiSelecting a subset of places from the other places as a load sample; thus, for shared sites li，P(li|u，ti，C(u，li) 1 for the location of negative sampling
Thus, a likelihood calculation formula for the data set containing the positive and negative examples is derived:
where u is the user and liThe location of the ith location in the geographic information sharing track of the user u, namely the given location,
is aiNegative sampling of (t)iIs the user shares liTime of geographic information, C (u, l)i) Is user u is sharing liGeographical information of places visited before and after. The final optimization target is maximum likelihood estimation, namely a calculation formula of maximum likelihood; wherein, P (l) before the specification is giveni|u，ti，C(u，li) Post visit location C (u, l)i) In case of (2), user u is at tiTime of day access liThe conditional probability of a place is calculated by a deep learning model;
as shown in fig. 2, in this deep learning model, a user U is represented as a user vector by a user embedding matrix U, a given place and places visited before and after are respectively represented as a place vector by a place embedding matrix L, the place vectors visited before and after are averaged and then spliced with the user vector, the geographical position sharing time and the vector of the given place, and then the place C (U, L) visited before and after is given is obtained by a Sigmoid functioni) In case of (2), user u is at tiTime of day access liConditional probability of a place P (l)i|u，ti，C(u，li) Or negative sampling site
Conditional probability of (2)
And (3) the binary cross entropy is used as a loss function, the likelihood is maximized by utilizing a gradient descent method, and a user embedding matrix U and a place embedding matrix L are estimated, so that the user embedding matrix contains the representation of the user geographical position sharing track and is used for a next user matching degree calculation system.
The user matching degree calculation system: in this step, the matching degree of each pair of users is calculated through a deep neural network model, as shown in fig. 3; the input of the model is two users, and the embedded vector representation of the two users is respectively calculated through a user embedded matrix U; then splicing the two vectors, and calling the spliced vector as H0Above this, q layers of fully connected neural network H are connected1，……，Hq，HqI.e. an implicit expression of the relationship between two users, in the q-layer fully-connected neural network, a ReLU function is used as an activation function of the fully-connected neural networks, namely Hk＝ReLU(WkHk-1+bk) (ii) a And finally, calculating the friend matching degrees of the two users by using a Sigmoid function, namely the friend matching degree Fij＝sigmoid(Wq+1Hq+bq+1)。
Only the sample in the user's buddy relationship data is reliable because there may be a situation where two users have a high degree of matching but have not yet become buddies. Thus, a negative sampling technique is also used in this step to randomly select a subset from the user pairs other than the known buddy relationship as the negative sample data set. If the two users are friends, the matching degree of the two users is 1; if a pair of users is in the negative examples dataset, then their degree of match is 0. And then, training the deep neural network model by using the matching degree data set containing the positive and negative samples, wherein the loss function is a binary cross entropy, the optimization method is a gradient descent method, and finally, the values of all coefficients in the model are determined.
The friend recommendation system comprises: and recommending m friends for each user, and for one user u, sorting the users which are not friends of the user according to the matching degree, selecting the first m users from the users, and recommending the users to the user u.
The above description is only an example of the present invention, and is not intended to limit the present invention. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims (5)
1. A social network friend recommendation system based on a geographic position sharing track is characterized in that: the friend recommending method comprises a data collecting system, a user characteristic representing system, a user matching degree calculating system and a friend recommending system, wherein the data collecting system is used for collecting user information, the user information is distinguished and represented according to the user characteristic representing system, simulation calculation is carried out according to the user matching degree, and the friend recommending system is given to a user according to a calculation result, wherein in the step of the user characteristic representing system, the front and back sequence of user geographical position sharing, namely a geographical position sharing track, is integrated into the user characteristic representing system.
2. The system of claim 1, wherein the social network friend recommendation system based on the geographical location sharing track comprises: the data collection system collects the geographic position sharing information and social information of the social users through social softwareClimb out Source dataThe geographical location sharing information comprises sharing time, geographical location longitude and latitude, location merchant names, location merchant types and the like, and the social information comprises a friend list of each user and friend adding time.
3. The system of claim 1, wherein the social network friend recommendation system based on the geographical location sharing track comprises: the user characteristic representation system adopts an unsupervised learning method, each user and each place are respectively represented as a vector according to the characteristics of the user geographical position sharing track, the sharing time and the like, and the theoretical basis of the model is that the geographical position shared by the users can pass through the user sharing habit and the places before and after the geographical position.
4. The system of claim 1, wherein the social network friend recommendation system based on the geographical location sharing track comprises: the user matching degree calculation system is internally provided with a deep neural network model, the matching degree of each pair of users is calculated through the deep neural network model, the model inputs two users, and the embedded vector representations of the two users are respectively calculated through a user embedded matrix U; then splicing the two vectors, and calling the spliced vector as H0Above this, q layers of fully connected neural network H are connected1，……，Hq，HqI.e. an implicit expression of the relationship between two users, in the q-layer fully-connected neural network, a ReLU function is used as an activation function of the fully-connected neural networks, namely Hk＝ReLU(WkHk-1+bk) (ii) a And finally, calculating the friend matching degrees of the two users by using a Sigmoid function, namely the friend matching degree Fij＝sigmoid(Wq+1Hq+bq+1)。
5. The system of claim 1, wherein the social network friend recommendation system based on the geographical location sharing track comprises: the friend recommendation system recommends m friends for each user, and for one user u, sorts the users who are not friends of the user according to the matching degree, selects the first m users from the users, and recommends the users to the user u.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110314101.6A CN112948711A (en) | 2021-03-24 | 2021-03-24 | Social network friend recommendation system based on geographical position sharing track |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110314101.6A CN112948711A (en) | 2021-03-24 | 2021-03-24 | Social network friend recommendation system based on geographical position sharing track |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112948711A true CN112948711A (en) | 2021-06-11 |
Family
ID=76228105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110314101.6A Pending CN112948711A (en) | 2021-03-24 | 2021-03-24 | Social network friend recommendation system based on geographical position sharing track |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112948711A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130041862A1 (en) * | 2010-04-23 | 2013-02-14 | Thomson Loicensing | Method and system for providing recommendations in a social network |
| CN104021233A (en) * | 2014-06-30 | 2014-09-03 | 电子科技大学 | Social network friend recommendation method based on community discovery |
| CN104881459A (en) * | 2015-05-22 | 2015-09-02 | 电子科技大学 | Friend recommendation method of mobile social network |
| CN105471715A (en) * | 2015-12-14 | 2016-04-06 | 西安交通大学 | Mobile social network friend recommending method and system based on geographical position change |
-
2021
- 2021-03-24 CN CN202110314101.6A patent/CN112948711A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130041862A1 (en) * | 2010-04-23 | 2013-02-14 | Thomson Loicensing | Method and system for providing recommendations in a social network |
| CN104021233A (en) * | 2014-06-30 | 2014-09-03 | 电子科技大学 | Social network friend recommendation method based on community discovery |
| CN104881459A (en) * | 2015-05-22 | 2015-09-02 | 电子科技大学 | Friend recommendation method of mobile social network |
| CN105471715A (en) * | 2015-12-14 | 2016-04-06 | 西安交通大学 | Mobile social network friend recommending method and system based on geographical position change |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hang et al. | Exploring student check-in behavior for improved point-of-interest prediction | |
| Xu et al. | A survey for mobility big data analytics for geolocation prediction | |
| Sadilek et al. | Finding your friends and following them to where you are | |
| CN109726336B (en) | POI recommendation method combining travel interest and social preference | |
| Barchiesi et al. | Modelling human mobility patterns using photographic data shared online | |
| Al-Molegi et al. | Move, attend and predict: An attention-based neural model for people’s movement prediction | |
| CN106022934A (en) | Potential friend discovering method based on moving trajectory pattern and system | |
| CN102945258A (en) | Method and system for obtaining reliability of users in social networking services | |
| CN111241419A (en) | Next interest point recommendation method based on user relationship embedded model | |
| CN113254999A (en) | User community mining method and system based on differential privacy | |
| CN113139140A (en) | Tourist attraction recommendation method based on space-time perception GRU and combined with user relation preference | |
| Ju et al. | Relationship strength estimation based on Wechat Friends Circle | |
| CN111738447B (en) | Mobile social network user relationship inference method based on spatio-temporal relationship learning | |
| CN111695019A (en) | Method and device for identifying associated account | |
| Kang et al. | How urban places are visited by social groups? Evidence from matrix factorization on mobile phone data | |
| Pang et al. | Intercity simulation of human mobility at rare events via reinforcement learning | |
| Bozanta et al. | Developing a contextually personalized hybrid recommender system | |
| CN112948711A (en) | Social network friend recommendation system based on geographical position sharing track | |
| CN110781256A (en) | Method and device for determining POI (Point of interest) matched with Wi-Fi (Wireless Fidelity) based on transmitted position data | |
| Zhang et al. | GEVR: An event venue recommendation system for groups of mobile users | |
| Chen et al. | Multi-context embedding based personalized place semantics recognition | |
| Jenson et al. | Mining location information from users' spatio-temporal data | |
| CN114048380A (en) | Interest point recommendation method based on graph neural network | |
| Yoshikawa et al. | A fake news dissemination model based on updating reliability and doubt among individuals | |
| Ding et al. | Modeling personalized out-of-town distances in location recommendation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210611 |
|
| RJ01 | Rejection of invention patent application after publication |