CN105141499A - Social network relationship recommendation method based on privacy degree and publicity degree - Google Patents

Abstract

The invention discloses a social network relationship recommendation method based on the privacy degree and the publicity degree. According to the invention, through defining the privacy degree and the publicity degree of social network relationship, measurement of the relationship between people is analyzed and calculated, the intimacy degree of the relationship is judged, and new social relationship is recommended and established intelligently for users, thereby effectively digging potential relationship between people in the social contact field, enriching the social relationship network, and improving the social network bonding degree of the users. The social network relationship recommendation method disclosed by the invention is relatively simple in process and good in effect.

Description

A kind of social network relationships recommend method based on secret degree and known degree

Technical field

The invention belongs to Artificial technical field of intelligence, be specifically related to a kind of design of the social network relationships recommend method based on secret degree and known degree.

Background technology

Along with the development of the Internet, people combine increasing for daily routines with the Internet, and the service for human social activity starts more and more by everybody concern, and current social networks has become one of important base application of the network user.The diversified network application that social networks (SocialNetworkService, SNS) brings is changing the individual use habit to the Internet, plays positive impetus to the development of the Internet simultaneously.Social networks, while safeguarding personal relationship, has also expanded new relation further.Along with the continuous expansion of SNS userbase, real-life interpersonal relationships extends in network by increasing user, based on society relation, simulate or rebuild the human relation network of society, individual work, emotion are closely connected with actual life, the part become a reality in life.Social network sites has very large potentiality fundamentally changing in the social life such as human communication and community activity mode.

Along with the development of network technology, increasing social networking service is providing the platform of communication for users, but how effectively to expand customer relationship, intelligence is excavated, is safeguarded and support interpersonal relation, understanding the relational network of user intelligently, is the key problem that each SNS must solve.Although more existing ralation methods based on shortest path, graph theory at present, but these algorithms or function more weak, relation excavation dynamics is little, or complexity is too high, when number of users is larger, the speed of service is extremely slow, well can not adapt to current social networking service present situation and requirement.

Relational links

In social networks, interpersonal relation is that the mode linked exists.Comprise " weak link " and " linking by force " two kinds of forms.What wherein weak link embodied is the process of information flow, is inter-trade cross-cutting Information Communication; " link by force " and then reflect everyone the most intimate relation at one's side.The resource of interpersonal relationship, by the mode of the Internet, can be excavated out by SNS completely.

Here both contained " linking by force " that those are got close to, also contains " weak link " that those are not met for a long time.By SNS, user can be familiar with like a cork " friend of friend ", thus by the people that the people of understanding finds oneself to need, expands the human connection of oneself.Meanwhile, user also scientifically can manage oneself interpersonal relationships net resource by this platform of SNS, for oneself winning more chance.The value root of SNS is just that the authenticity of this platform information, user provide oneself actual data, and entire society's network completely based on realistic individual thing and relation, thus provides true, credible, an effective social stage.How on the social stage that this is virtual, developing valuable application, promote the emotion between friend and information interchange effectively, is the key playing SNS value.

Six Degrees is theoretical

" Six Degrees theory is " theoretical also known as making six degrees of separation (SixDegreesofSeparation).This theory can generically be illustrated as: " between you and any one stranger, the people at institute interval can not more than six, and that is, by six people, you just can be familiar with any one stranger at most." this theory results from the sixties in 20th century, proposed by American Psychologist Mil Glenn.

This theory is thought, people by six layers of interpersonal relationships just can find tellurian anyone.Although it still only rests on controversial " hypothesis " stage so far, cause research and the concern of every field scholar.

Restlet framework

The issue of the relation of social networks presents on network with the form of " service ", therefore needs to adopt service framework to issue out in the mode of WebService by SNS.Restlet is the lightweight REST framework under a Java, boundary between Web site and Web service that it is fuzzy, thus helps developer to build Web application.

REST is a kind of client terminal/server structure, its connection protocol has Stateless, require that the information that client transmits through stateless connection protocol at every turn must comprise state informations all in application, namely each request from client to server must comprise understands the necessary all information of this request, can not utilize any storage context on the server, therefore session status all will be kept at client.The Stateless of REST improves the observability of system, reliability and scalability, the state between multiple request need not be preserved, server component just also can simplify its realization by releasing resource rapidly further, simultaneously surveillance also need not check the data of multiple request in order to determine whole character of asking, in addition, the Stateless of communication itself can allow the difference request in a series of request of the process of different servers, improves the autgmentability of server.But but reduce the performance of network like this, because client has to send some data repeated, so the noticeable performance of efficiency and user in order to improve system, and make system have stratification, REST employs caching mechanism.

Caching component plays the part of an arbitrator between clients and servers, and the response of previous request can be reused, and to respond same request after a while, if by this request forward to server, the response obtained may be identical with response existing in buffer memory.But equally there is a problem in this, if data outmoded in buffer memory exactly with will ask to be dealt into data differences that server obtains very greatly, will reduce reliability, key is the cache policy selected.This problem of performance is the perfect solution of neither one inherently, can only find a best balance point according to the needs of system as far as possible.

Servlet framework

The customer relationship of social networks is issued result and is existed with the form of Web page.Servlet is the intermediate layer that client asks to respond with server, is the java application of the server end being positioned at Web server inside, has the characteristic independent of platform and agreement, can generate dynamic Web page.Be different from the java application of traditional start up with command-line options, Servlet is loaded by Web server, and this Web server must comprise the Java Virtual Machine supporting Servlet.From realization, Servlet can respond the request of any type, but Servlet is only used for expanding the Web server based on http protocol in most cases.

Secret degree

Secret degree describes the far and near close and distant degree of relation between user and good friend, secret degree is higher, illustrate that user is nearer to friend relation, then larger from the reliability of this good friend's obtaining information, meanwhile, friend recommendation webpage, the article delivered, to be concerned to the answer of problem the chance adopted also larger.Therefore the people that those secret degree are high, should be in more forward position in relation is recommended.

Known degree

Known degree is the degree representing that a people is known by the public, understands, and is the breadth and depth of social influence, is the objective yardstick evaluating reputation size.Known degree and his personal story of a people have close relationship, and such as one had more than ten years experience and the known degree being responsible for the expert of multinomial catenet architecture design obviously can be greater than at the beginning of one and relates to the people that this field there is not experience at computer network field.The known degree of a people is higher, and the webpage that he recommends, the blog delivered, the authority of answering a question are higher, should be in higher priority in relation recommendation.

Summary of the invention

The object of the invention is in order to solve in prior art based on the ralation method of shortest path, graph theory or function more weak, relation excavation dynamics is little, or complexity is too high, when number of users is larger, the speed of service is extremely slow, well can not adapt to current social networking service present situation and the problem of requirement, propose a kind of social network relationships recommend method based on secret degree and known degree.

Technical scheme of the present invention is: a kind of social network relationships recommend method based on secret degree and known degree, comprises the following steps:

The direct good friend of S1, acquisition good friend requestor A gathers F _a={ U ₁, U ₂..., U _m;

The commending friends set FR of final good friend requestor A is set up in S2, initialization _a;

S3, friend recommendation threshold value M is set;

S4, calculating F _ain the relation weights of each direct good friend and A;

S5, search the good friend U with the relation maximum weight of A _k, obtain U _kdirect good friend set

S6, definition U _kdirect good friend set secret degree dimension values for A is two degree of good friend's set of 2, calculates in the relation weights of each two degree of good friends and A;

S7, all two degree of good friends being more than or equal to M with the relation weights of A are added FR _a;

S8, newly FR is added to each _acommending friends V _j, obtain its direct good friend's set

S9, the secret degree dimension values of good friend is added 1, each good friend in the N dimension good friend set of each A that calculation procedure S8 obtains and the relation weights of A;

S10, all N dimension good friends being more than or equal to M with the relation weights of A are added FR _a;

S11, judge whether good friend's secret degree dimension values equals 6;

If then enter step S12;

Then return step S8 if not;

S12, by commending friends set FR _arecommend good friend requestor A.

Further, commending friends set FR _abe initialized as empty set.

Further, commending friends set

Further, step S4 is specially:

To the direct good friend U of each A _i∈ F _a, calculate U according to formula (1) _iwith the relation weights of A

$W_{{\mathrm{AU}}_i}={\mathrm{\ρI}}_{{\mathrm{AU}}_i}+(1-\mathrm{\ρ})R_A---\left(1\right)$

In formula represent A and U _isecret degree, span is [0,1]; R _arepresent the maximum known degree in the direct good friend of A, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l].

Further, step S6 is specially:

Definition U _kdirect good friend set secret degree dimension values for A is two degree of good friend's set of 2, to each U _kdirect good friend, i.e. two degree of good friends of A v is calculated according to formula (2) _iwith the relation weights of A

$W_{{\mathrm{AV}}_i}={\mathrm{\ρI}}_{{\mathrm{AV}}_i}+(1-\mathrm{\ρ})R_{U_k}-{\mathrm{\σD}}_{{\mathrm{AV}}_i}---\left(2\right)$

In formula represent A and V _isecret degree, span is [0,1]; represent U _kdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l]; represent A and V _ithe dimension of good friend's secret degree, in this step σ represents the degree of concern of user to internuncial number, span be (0, l).

Further, step S9 is specially:

The secret degree dimension values of good friend is added 1, according to each good friend in the N dimension good friend set of each A of formula (3) calculation procedure S8 acquisition with the relation weights of A

$W_{{\mathrm{AX}}_i}={\mathrm{\ρI}}_{{\mathrm{AX}}_i}+(1-\mathrm{\ρ})R_{V_j}-{\mathrm{\σD}}_{{\mathrm{AX}}_i}---\left(3\right)$

In formula represent A and X _isecret degree, span is [0,1]; represent V _jdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l]; represent A and X _ithe dimension of good friend's secret degree, often perform a deuterzooid step, value add 1; σ represents the degree of concern of user to internuncial number, span be (0, l).

The invention has the beneficial effects as follows: the present invention is by defining the secret degree of social network relationships and known degree, analyze, calculate the tolerance of interpersonal relationship, the intimate degree of judgement relation, and accordingly for user's intelligent recommendation sets up new social networks, effectively excavate human relationship potential in social field, enriched social networks network, improve the social networks degree of adhesion of user, process is relatively simple, and effect is better.

Accompanying drawing explanation

Fig. 1 is a kind of social network relationships recommend method flow chart based on secret degree and known degree provided by the invention.

Embodiment

Below in conjunction with accompanying drawing, embodiments of the invention are further described.

In the present invention, if good friend requestor is A, we need for requestor A finds a suitable user B as its commending friends, make A and B meet three conditions below:

(1) relation of A and B is as far as possible intimate, then A obtains the possibility of useful information more greatly from B, and we use secret degree to weigh the intimate degree between A and B here.

Secret degree describes the far and near close and distant degree of relation between user and good friend, secret degree is higher, illustrate that user is nearer to friend relation, then larger from the reliability of this good friend's obtaining information, meanwhile, friend recommendation webpage, the article delivered, to be concerned to the answer of problem the chance adopted also larger.Therefore the people that those secret degree are high, should be in more forward position in relation is recommended.

(2) the known degree of targeted customer B should be high as far as possible, then the credibility of information that obtains from B of A is higher.

Known degree is the degree representing that a people is known by the public, understands, and is the breadth and depth of social influence, is the objective yardstick evaluating reputation size.Known degree and his personal story of a people have close relationship, and such as one had more than ten years experience and the known degree being responsible for the expert of multinomial catenet architecture design obviously can be greater than at the beginning of one and relates to the people that this field there is not experience at computer network field.The known degree of a people is higher, and the webpage that he recommends, the blog delivered, the authority of answering a question are higher, should be in higher priority in relation recommendation.

(3) from A to B, the go-between of experience should be the least possible, because often through a sponsor, secret degree will be decayed once, B helps the wish of A also can reduce gradually.

The invention provides a kind of social network relationships recommend method based on secret degree and known degree, as shown in Figure 1, comprise the following steps:

The direct good friend of S1, acquisition good friend requestor A gathers F _a={ U ₁, U ₂..., U _m;

Here the direct good friend of A is the once good friend that the secret degree dimension values of A is 1.

The commending friends set FR of final good friend requestor A is set up in S2, initialization _a;

Commending friends set FR _abe initialized as empty set, the commending friends of final good friend requestor A all can deposit in FR _ain.

Because the direct good friend of A gathers F _ain user be the good friend of A, there is no need to recommend A again, therefore FR _ain can not deposit F _ain user, namely

S3, friend recommendation threshold value M is set;

S4, calculating F _ain the relation weights of each direct good friend and A;

To the direct good friend U of each A _i∈ F _a, calculate U according to formula (1) _iwith the relation weights of A

$W_{{\mathrm{AU}}_i}={\mathrm{\ρI}}_{{\mathrm{AU}}_i}+(1-\mathrm{\ρ})R_A---\left(1\right)$

In formula represent A and U _isecret degree, span is [0,1].

Here the secret degree between user is described, I with variable I _aBnamely represent the secret degree of user A and B, close to 0, its value more represents that the secret degree of A and B is lower, otherwise then secret degree is higher.If A and B is not familiar with completely, then I _aB=0, and if only if A and B identical time I _aB=1.In addition, the situation that B and B are not familiar with A is familiar with, so I due to A may be there is _aB≠ I _bA.When initial, I _aBset by A, automatically upgraded according to the activity between A, B and operation by system later.

R _arepresent the maximum known degree in the direct good friend of A, span is [0,1].

If variable C _awith representing the known degree of user A in setting field, such as art of mathematics, physics field etc., span is [0,1].Close to 0, this value more represents that the known degree of user is lower, otherwise then the known degree of user is higher.C _ainitial value set by management staff or audit crew according to the natural quality of this user A by system, upgraded according to the activity of user and operation by system later.Due to F _a={ U ₁, U ₂..., U _m, definition then R _arepresent maximum known degree in the direct good friend of A.

ρ represents ? in shared importance ratio, span is [0, l].The less explanation of ρ value is right concern fewer, and to R _apay close attention to more.When ρ=0, represent that in the direct good friend of user A, maximum known degree is unique factor paid close attention to, this algorithm is degenerated to maximum known degree in the direct good friend calculating user A; When ρ=1, represent A and U _isecret degree be unique factor paid close attention to, this algorithm is degenerated to and calculates A and U _isecret degree.

S5, search the good friend U with the relation maximum weight of A _k, obtain U _kdirect good friend set

S6, definition U _kdirect good friend set secret degree dimension values for A is two degree of good friend's set of 2, calculates in the relation weights of each two degree of good friends and A;

To each U _kdirect good friend, i.e. two degree of good friends of A v is calculated according to formula (2) _iwith the relation weights of A

$W_{{\mathrm{AV}}_i}={\mathrm{\ρI}}_{{\mathrm{AV}}_i}+(1-\mathrm{\ρ})R_{U_k}-{\mathrm{\σD}}_{{\mathrm{AV}}_i}---\left(2\right)$

In formula represent A and V _isecret degree, span is [0,1]; represent U _kdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l].

represent A and V _ithe dimension of good friend's secret degree, due to U in this step _kthe direct good friend of A, V _iu _kdirect good friend, A and V _irelation be pass through U _kcarry out transmitting, therefore claim V _ifor two degree of good friends of A, namely

σ represents the degree of concern of user to internuncial number, and span is that (0, l), σ value is larger, weights decay faster.

S7, all two degree of good friends being more than or equal to M with the relation weights of A are added FR _a;

S8, newly FR is added to each _acommending friends V _j, obtain its direct good friend's set

S9, the secret degree dimension values of good friend is added 1, each good friend in the N dimension good friend set of each A that calculation procedure S8 obtains and the relation weights of A;

Here according to each good friend in the N dimension good friend set of each A of formula (3) calculation procedure S8 acquisition with the relation weights of A

$W_{{\mathrm{AX}}_i}={\mathrm{\ρI}}_{{\mathrm{AX}}_i}+(1-\mathrm{\ρ})R_{V_j}-{\mathrm{\σD}}_{{\mathrm{AX}}_i}---\left(3\right)$

In formula represent A and X _isecret degree, span is [0,1]; represent V _jdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l]; represent A and X _ithe dimension of good friend's secret degree, often perform a deuterzooid step, value add 1; σ represents the degree of concern of user to internuncial number, span be (0, l).

S10, all N dimension good friends being more than or equal to M with the relation weights of A are added FR _a;

S11, judge whether good friend's secret degree dimension values equals 6, if then enter step S12, then returns step S8 if not;

Here to get the secret degree dimension values upper limit be 6 is the results drawn according to Six Degrees theory.

S12, by commending friends set FR _arecommend good friend requestor A.

Before meeting, namely the suitable user B of described three conditions is contained in commending friends set FR _ain.

Those of ordinary skill in the art will appreciate that, embodiment described here is to help reader understanding's principle of the present invention, should be understood to that protection scope of the present invention is not limited to so special statement and embodiment.Those of ordinary skill in the art can make various other various concrete distortion and combination of not departing from essence of the present invention according to these technology enlightenment disclosed by the invention, and these distortion and combination are still in protection scope of the present invention.

Claims (6)

1., based on a social network relationships recommend method for secret degree and known degree, it is characterized in that, comprise the following steps:

The direct good friend of S1, acquisition good friend requestor A gathers F _a={ U ₁, U ₂..., U _m;

The commending friends set FR of final good friend requestor A is set up in S2, initialization _a;

S3, friend recommendation threshold value M is set;

S4, calculating F _ain the relation weights of each direct good friend and A;

S5, search the good friend U with the relation maximum weight of A _k, obtain U _kdirect good friend set

S6, definition U _kdirect good friend set secret degree dimension values for A is two degree of good friend's set of 2, calculates in the relation weights of each two degree of good friends and A;

S7, all two degree of good friends being more than or equal to M with the relation weights of A are added FR _a;

S8, newly FR is added to each _acommending friends V _j, obtain its direct good friend's set

S9, the secret degree dimension values of good friend is added 1, each good friend in the N dimension good friend set of each A that calculation procedure S8 obtains and the relation weights of A;

S10, all N dimension good friends being more than or equal to M with the relation weights of A are added FR _a;

S11, judge whether good friend's secret degree dimension values equals 6;

If then enter step S12;

Then return step S8 if not;

S12, by commending friends set FR _arecommend good friend requestor A.

2. the social network relationships recommend method based on secret degree and known degree according to claim 1, is characterized in that, described commending friends set FR _abe initialized as empty set.

3. the social network relationships recommend method based on secret degree and known degree according to claim 1, is characterized in that, described commending friends set

4. the social network relationships recommend method based on secret degree and known degree according to claim 1, it is characterized in that, described step S4 is specially:

To the direct good friend U of each A _i∈ F _a, calculate U according to formula (1) _iwith the relation weights of A

$W_{{\mathrm{AU}}_i}={\mathrm{\ρI}}_{{\mathrm{AU}}_i}+(1-\mathrm{\ρ})R_A---\left(1\right)$

In formula represent A and U _isecret degree, span is [0,1]; R _arepresent the maximum known degree in the direct good friend of A, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l].

5. the social network relationships recommend method based on secret degree and known degree according to claim 1, it is characterized in that, described step S6 is specially:

Definition U _kdirect good friend set secret degree dimension values for A is two degree of good friend's set of 2, to each U _kdirect good friend, i.e. two degree of good friends of A v is calculated according to formula (2) _iwith the relation weights of A

$W_{{\mathrm{AV}}_i}={\mathrm{\ρI}}_{{\mathrm{AV}}_i}+(1-\mathrm{\ρ})R_{U_k}-{\mathrm{\σD}}_{{\mathrm{AV}}_i}---\left(2\right)$

In formula represent A and V _isecret degree, span is [0,1]; represent U _kdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l]; represent A and V _ithe dimension of good friend's secret degree, in this step σ represents the degree of concern of user to internuncial number, and span is (0,1).

6. the social network relationships recommend method based on secret degree and known degree according to claim 1, it is characterized in that, described step S9 is specially:

The secret degree dimension values of good friend is added 1, according to each good friend in the N dimension good friend set of each A of formula (3) calculation procedure S8 acquisition with the relation weights of A

$W_{{\mathrm{AX}}_i}={\mathrm{\ρI}}_{{\mathrm{AX}}_i}+(1-\mathrm{\ρ})R_{V_j}-{\mathrm{\σD}}_{{\mathrm{AX}}_i}---\left(3\right)$

In formula represent A and X _isecret degree, span is [0,1]; represent V _jdirect good friend in maximum known degree, span is [0,1]; ρ represents ? in shared importance ratio, span is [0, l]; represent A and X _ithe dimension of good friend's secret degree, often perform a deuterzooid step, value add 1; σ represents the degree of concern of user to internuncial number, span be (0, l).