CN107506455A - A kind of friend recommendation method for merging user social contact circle similarity - Google Patents
A kind of friend recommendation method for merging user social contact circle similarity Download PDFInfo
- Publication number
- CN107506455A CN107506455A CN201710755695.8A CN201710755695A CN107506455A CN 107506455 A CN107506455 A CN 107506455A CN 201710755695 A CN201710755695 A CN 201710755695A CN 107506455 A CN107506455 A CN 107506455A
- Authority
- CN
- China
- Prior art keywords
- user
- msub
- mrow
- similarity
- circle
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000006116 polymerization reaction Methods 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000011524 similarity measure Methods 0.000 abstract description 12
- 238000004422 calculation algorithm Methods 0.000 abstract description 8
- 238000007418 data mining Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 8
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 230000002776 aggregation Effects 0.000 description 3
- 238000004220 aggregation Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 206010034719 Personality change Diseases 0.000 description 1
- 230000018199 S phase Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005295 random walk Methods 0.000 description 1
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
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention proposes a kind of friend recommendation method for merging user social contact circle similarity, is related to the identification of social networks social circle, social circle's Similarity measures and user's Similarity measures, belongs to data mining and commending system field.A kind of method is proposed, using clustering method, corporations' division is carried out to user good friend using Fast Unfolding algorithms and then calculates corporations' similarity between user, Top n is selected, further according to user property, calculates the similarity between user.By the way that user social contact circle similarity and user's similarity are combined, recommendation list is calculated, realizes high-precision friend recommendation.
Description
Technical field
The present invention relates to data mining and commending system field, more particularly to a kind of fusion user social contact circle similarity
Friend recommendation method.
Background technology
With Web2.0 development, social networks turns into the typical case occurred in Web2.0 mode developments, individual character
Change recommended technology to be also widely used in social networks.Social networks that is to say social network service (Social
Network Service, SNS), intuitively say and can be described as network social activity, its development is broadly divided into:E-mail,
BBS, BLOG, Facebook/ Renren Network etc. stage.In social networks, user is by adding the mode such as good friend, concern mechanism
Carry out interaction and Information Sharing, and adding good friend between each other necessarily makes to produce a kind of contact between user, with when
Between development, a kind of good friend's social networks is together form between user and its good friend and the good friend of good friend, therefrom we can
To recognize the hobby of user, Information Sharing, communicational aspects and its active degree etc..Further, since social networks
In user base number it is bigger, and the user in network is the stranger for being short in understanding and trusting between each other mostly, therefore
Friend relation can not possibly be arbitrarily formed, these are all the problem of carrying out having to consider during friend recommendation in social networks.
It is various to push away constantly by the concern of scholars as the core component of commending system, the research of proposed algorithm
Algorithm is recommended to emerge in an endless stream, but proposed algorithm has certain limitation, and what is recommended is the object with " physical property " mostly, and
Research in terms of recommendation about social network user or good friend is then relatively fewer.Even if in existing friend recommendation algorithm,
It is that structure good friend's recommended models such as information, user's topic preference are chained by certain method such as user mostly, two users get over
Similar, they are more likely to become good friend, main to include using Saltonx similarity factors, Jaccard similarity factors as representative
Based on the similarity indices of common neighbours, the similarity indices based on node similarity using Adamic-Adar as representative;Together
When, in addition to the SimRank index scheduling algorithms based on random walk.As can be seen here, traditional user recommends method mainly with individual
Recommended based on body similarity, few influence for considering community's factor.
The content of the invention
In order to overcome defect present in above-mentioned prior art, it is an object of the invention to provide one kind to merge user social contact circle
The friend recommendation method of similarity.This method is produced initially by the way that social circle's similarity is identified, calculated to user social contact circle
Recommendation list, the consequently recommended list of user's similarity generation is calculated, while with the information of user social contact circle and user itself letter
Breath, which improves, recommends precision.
In order to realize the above-mentioned purpose of the present invention, the invention provides a kind of good friend for merging user social contact circle similarity to push away
Method is recommended, is comprised the following steps:
S1:Gather user profile;
S2:Using the user profile gathered in S1, user social contact circle is identified;
S3:User social contact circle similarity is calculated, and carries out handing over circle sequencing of similarity;
S4:Using the friendship circle sequencing of similarity obtained in S3, with reference to user property, user's similarity is calculated;
S5:The user's similarity obtained to S4 will be ranked up, and obtain the friend recommendation list of corresponding user.
Because user base number can elapse over time in social networks, become quite huge, and user recommends to require
Certain real-time.In order to be quickly obtained user social contact circle similarity ranking, Fast Unfolding schemes are selected in S2.Mould
Lumpiness (modularity) refers to connecting the ratio shared by the side of community structure internal vertex in network, subtracts in same society
The desired value of the ratio of the two nodes is arbitrarily connected under unity structure.The excellent of such division can be portrayed by modularity
Bad, modularity is bigger, then the effect of social circle's division is better, and its formula of the calculating of modularity is:
Wherein, Ai,jWhat is represented is weight between summit i and summit j, and m is company's side number of network, Pi,jRepresent be with
The desired value for the side number that summit i is connected with summit j, ciWhat is represented is community that summit is assigned to, δ (ci,cj) be used to judge top
Whether point i and summit j is divided in same community, if so, then returning to 1, otherwise, returns to 0.
User social contact circle is identified in above-mentioned S2 steps and comprised the following steps:
S21:Using each user as a node, node is initialized, a user is represented with a node, represents to use with side
The relation at family, there is contact between user, being then considered as has relation;The weight on side represents the number of contact.
S22:By in the community where each node division to node adjacent thereto, computing module degree Q so that modularity
Q reaches maximum.
S23:The community divided in S22 polymerization is turned into a point, network is reconfigured according to the community structure of generation.
S24:The step of repeating S22, S23, untill the structure in network no longer changes, identifies the social activity of user
Circle.
The formula of calculating user social contact circle similarity is in above-mentioned steps S3
Wherein pijRepresent user u1、u2The number of overlapping social circle's interior joint, eijRepresent user u1、u2In overlapping social circle
The bar number on overlapping side;By user u1Social circle's recognition result be labeled as u1c1,u1c2,u1c3…u1cm, user u2Social circle know
Other result queue is u2c1,u2c2,u2c3…u2cn, the degree of overlapping of one overlapping social circle of user is pij×(eij+ 1), m represents to use
Family u1The social number of turns, n represents user u2The social number of turns.
Specifically, the user property in above-mentioned steps S4, including the base attribute of user, the position of user, user it is upper
Net preference information.
The base attribute of the user is expressed as
userInfo(u1)={ basic (u1),tag(u1)}
Wherein, basic (u1) represent user u1Personal information description, tag (u1) represent u1Label information;
User u1, u2Base attribute similarity is designated as
sim(userInfo(u1),userInfo(u2))=λ1sim(basic(u1),basic(u2))+λ2sim(tag
(u1),tag(u2))
Wherein λ1+λ2=1, basic (u2) represent u2The personal information description of user, tag (u2) represent u2Label information.
The information of the position of the user is expressed as place (u1), user u1, u2Location similarity is sim (place
(u1),place(u2));According to province, city, the similarities and differences for commonly using dwell point, sim (place (u are determined1),place(u2))
Value.
The online preference of the user is expressed as net (u1), user u1, u2Preference similarity of surfing the Net is sim (net (u1),
net(u2));User's cosine similarity is often calculated with app information, online duration, flow information according to user, represents that user's is upper
Net preference similarity.
The formula of user's similarity is
Relationship of Coefficients is β1+β2+β3=1.
Beneficial effect:
The present invention carries out corporations' division to user good friend using Fast Unfolding algorithms and then calculated between user
Corporations' similarity, by the way that the Similarity Measure of good friend is divided into two steps:User social contact circle Similarity Measure and user's phase
Calculated like degree, so as to solve in conventional recommendation algorithm, calculate similarity commending friends only by user property, recommend precision
The problem of not high.
Brief description of the drawings
The above-mentioned and/or additional aspect and advantage of the present invention, with reference to accompanying drawings below to that will become in the description of embodiment
Substantially and it is readily appreciated that, wherein:
Fig. 1 is the overall flow structural representation of the present invention;
Fig. 2 is the user social contact circle identification process figure of the present invention;
Fig. 3 is the user social contact circle Similarity Measure schematic flow sheet of the present invention;
Fig. 4 is user's Similarity Measure schematic flow sheet of the present invention.
Embodiment
The specific implementation to the present invention is further explained in detail below in conjunction with the accompanying drawings.
The invention provides a kind of friend recommendation method for merging user social contact circle similarity, comprise the following steps:
S1:Each node is initialized, by the community where each node division to node adjacent thereto, so that
Modularity (Modularity) Q reaches maximum, and this stage is referred to as Modularity Optimization (modularity optimization).
S2:S1 is marked off to the community's polymerization come turns into a point, i.e., the communities of users structure weight generated according to previous step
Neotectonics network.The process of the above is repeated, untill the structure in network no longer changes.This process is referred to as Community
Aggregation (community's polymerization).
S3:By user u1Social circle's recognition result be labeled as u1c1,u1c2,u1c3…u1cm, user u2Social circle identification
Result queue is u2c1,u2c2,u2c3…u2cn, u2c1Represent user u2First social circle.pijRepresent user u1、u2Overlapping society
Hand over the number of circle interior joint, eijRepresent user u1、u2The bar number on overlapping side in overlapping social circle.Calculate one overlapping social activity of user
The degree of overlapping of circle is pij×(eij+ 1), according toCalculate social circle's similarity of user, m
Represent user u1The social number of turns, n represents user u2The social number of turns.Sort to obtain society from big to small according to social circle's similarity
Hand over circle similarity ranking Top-n good friends, the input as next step.
S4:Calculate the base attribute similarity of user.User u1Base attribute be expressed as userInfo (u1)={ basic
(u1),tag(u1), basic (u1) represent user u1Personal information description, tag (u1) represent u1Label information.User u1,
u2Base attribute similarity is designated as sim (userInfo (u1),userInfo(u2))=λ1sim(basic(u1),basic(u2))+
λ2sim(tag(u1),tag(u2));Wherein λ1+λ2=1, λ1, λ2Represent coefficient.
S5:Calculate the location similarity of user, user u1Positional information be expressed as place (u1), user u1, u2Position
Similarity is designated as sim (place (u1),place(u2)).According to province, city, the similarities and differences for commonly using dwell point, sim is determined
(place(u1),place(u2)) value, wherein choosing the most frequently used three of user registers addresses as conventional dwell points.
S6:Calculate online preference similarity, user u1Online preference be expressed as net (u1).User u1, u2Online preference
Similarity is designated as sim (net (u1),net(u2)).App information, online duration, flow information is often used to calculate more than user according to user
String similarity, represent the first-class preference similarity of user.It is normal as user wherein to choose 5 sections of APP of user's frequency of use highest
Use APP.
S7:Calculate user u1, u2User's similarityWherein
Relationship of Coefficients is λ1+λ2+λ3=1.
S8:User's similarity caused by previous step is ranked up according to order from high in the end, produces user's similarity
Ranking Top-n, obtain the friend recommendation list of corresponding user.
Above-mentioned steps are further described in detail below in conjunction with accompanying drawing.
Fig. 1 is the overall flow structural representation of the present invention.The present invention mainly includes three parts:User social contact circle is known
Not, user social contact circle Similarity Measure and user's Similarity Measure, by the Similarity Measure of end user, obtain ranking and produce
Raw user's recommendation list.
When carrying out the identification of user social contact circle, user social contact network is converted into graph structure, a user is with a node
Represent there was contact between user, being then considered as has relation, and the present invention represents the relation of user with side, and the weight on side represents contact
Number.Then by the community where each node division to the node being adjacent, to cause modularity (Modularity)
Value constantly become big, the first step is marked off to community (Community) polymerization come turns into a point, i.e., is generated according to previous step
Community structure reconfigure network.The process of the above is repeated, untill the structure in network no longer changes, just have identified
The social circle of user.According to the social circle of user, the overlapping node of user social contact circle and side are calculated, calculates user social contact circle
Similarity, produce social circle similarity ranking.Obtain social circle similarity highest Top-n user, from user's base attribute,
Positional information, online three dimensions of preference carry out the measurement of user's similitude, adjust the weight shared by three dimensions, are recommended
List.
Fig. 2 is the user social contact circle identification process figure of the present invention.Mainly include two processes:Modularity
Optimization and Community Aggregation.First stage is referred to as Modularity Optimization, mainly
It is by the community where each node division to the node being adjacent, to cause the value of modularity constantly to become big;Second-order
Section is referred to as Community Aggregation, and the first step is mainly marked off to the community's polymerization come turns into a point, i.e. basis
The community structure of previous step generation reconfigures network.The process of the above is repeated, untill the structure in network no longer changes.
Specific algorithmic procedure includes:
1. initialization, each point is divided in different social circles;
2. pair each node, by each social circle put where trial is divided into the point being adjacent, calculate now
Modularity, whether the difference △ Q for judging to divide front and rear modularity are positive number, if positive number, then receive this division, if not
For positive number, then this division is abandoned;
3. the process of the above is repeated, untill it can not increase modularity again;
Scheme 4. construction is new, what each point in new figure represented is that each social circle come is marked in step 3, continues executing with step
Rapid 2 and step 3, untill the structure of social circle no longer changes.
The calculation formula of modularity mentioned above is:Wherein, Ai,jRepresent
It is the weight between summit i and summit j, m is company's side number of network, Pi,jWhat is represented is the side number being connected with summit i with summit j
Desired value, ciWhat is represented is community that summit is assigned to, δ (ci,cj) be used to judge whether summit i is divided in summit j
In same community, if so, then returning to 1, otherwise, 0 is returned to.
Fig. 3 is the user social contact circle Similarity Measure schematic flow sheet of the present invention.The result of communities of users identification is as use
The input of family social circle Similarity Measure.First, data processing is carried out, by user u1Social circle's recognition result be labeled as u1c1,
u1c2,u1c3…u1cm, user u2Social circle's recognition result be labeled as u2c1,u2c2,u2c3…u2cn.It is overlapping to calculate user one
The degree of overlapping of social circle is pij×(eij+ 1), wherein pijRepresent user u1、u2The number of overlapping social circle's interior joint, eijRepresent
User u1、u2The bar number on overlapping side in overlapping social circle.According toCalculate the social circle of user
Similarity, so as to sort to obtain user social contact circle similarity ranking.
Fig. 4 is user's Similarity Measure schematic flow sheet of the present invention.Using user social contact circle ranking as input, from user
Base attribute, the position of user, three dimensions of online preference of user carry out the calculating of user's similarity.
First, the base attribute similarity of user is calculated.basic(u1) represent user u1Personal information description, tag
(u1) represent u1Label information.Personal information description includes user's sex, age, is used as use in net duration, three labels of selection
The representative of family label.User u1Base attribute be expressed as userInfo (u1)={ basic (u1),tag(u1), user u1, u2
Base attribute similarity is designated as sim (userInfo (u1),userInfo(u2))=λ1sim(basic(u1),basic(u2))+λ2sim(tag(u1),tag(u2));Wherein Relationship of Coefficients is λ1+λ2=1.The personal information description similarity of user, the present invention adopt
Calculated with edit distance approach, the method is used for the similarity for calculating two natural language sentences.Similarity span for [0,
1], formula is as follows:sim(basic(u1),basic(u2))=1-Distance (basic (u1),basic(u2))/max
(Length(basic(u1)),Length(basic(u2))), wherein, basic (u1)、basic(u2) editing distance be
Distance(basic(u1),basic(u2)), Length (x) represents x length.Similarly, user tag similarity is also using volume
Distance method is collected to calculate.
Calculate the location similarity of user, user u1Positional information be expressed as place (u1), user u1, u2Position is similar
Degree is designated as sim (place (u1),place(u2)).According to province, city, the similarities and differences for commonly using dwell point, sim (place are determined
(u1),place(u2)) value, wherein choosing the most frequently used three of user registers addresses as conventional dwell points;If province, city
City, conventional parked position all same, sim (place (u1),place(u2))=1;If province, city are identical, parked position is commonly used
Difference, sim (place (u1),place(u2))=2/3;If province is identical, city, conventional parked position are different, sim (place
(u1),place(u2))=1/3;If province, city, conventional parked position are different, sim (place (u1),place(u2))=0;
Calculate online preference similarity, user u1Online preference be expressed as net (u1).User u1, u2Preference of surfing the Net is similar
Degree is designated as sim (net (u1),net(u2)).User's cosine phase is often calculated with app information, online duration, flow information according to user
Like degree, the online preference similarity of user is represented.5 sections of APP of user's frequency of use highest are wherein chosen to commonly use as user
APP;
Finally, user u is calculated1, u2User's similarityWherein
Relationship of Coefficients is λ1+λ2+λ3=1.
Claims (10)
- A kind of 1. friend recommendation method for merging user social contact circle similarity, it is characterised in that comprise the following steps:S1:Gather user profile;S2:Using the user profile gathered in S1, user social contact circle is identified;S3:User social contact circle similarity is calculated, and carries out handing over circle sequencing of similarity;S4:Using the friendship circle sequencing of similarity obtained in S3, with reference to user property, user's similarity is calculated;S5:The user's similarity obtained to S4 will be ranked up, and obtain the friend recommendation list of corresponding user.
- A kind of 2. friend recommendation method for merging user social contact circle similarity according to claim 1, it is characterised in that:Institute State S2 and user social contact circle is identified and comprise the following steps:S21:Using each user as a node, node is initialized;S22:By in the community where each node division to node adjacent thereto, computing module degree Q so that modularity Q reaches To maximum;S23:The community divided in S22 polymerization is turned into a point, network is reconfigured according to the community structure of generation;S24:The step of repeating S22, S23, untill the structure in network no longer changes, identifies the social circle of user.
- A kind of 3. friend recommendation method for merging user social contact circle similarity according to claim 2, it is characterised in that:Institute State and node initialized in S21, be that a user is represented with a node, the relation of user is represented with side, had contact between user, Then being considered as has relation;The weight on side represents the number of contact.
- A kind of 4. friend recommendation method for merging user social contact circle similarity according to claim 2, it is characterised in that:Institute Stating the calculating of modularity in S22 its formula is:<mrow> <mi>Q</mi> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mn>2</mn> <mi>m</mi> </mrow> </mfrac> <munder> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> </munder> <mo>&lsqb;</mo> <msub> <mi>A</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> </msub> <mo>-</mo> <msub> <mi>P</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> </msub> <mo>&rsqb;</mo> <mi>&delta;</mi> <mrow> <mo>(</mo> <msub> <mi>C</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>C</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow>Wherein, Ai,jWhat is represented is weight between summit i and summit j, and m is company's side number of network, Pi,jRepresent be and summit i The desired value for the side number being connected with summit j, ciWhat is represented is community that summit is assigned to, δ (ci,cj) be used for judge summit i with Whether summit j is divided in same community, if so, then returning to 1, otherwise, returns to 0.
- A kind of 5. friend recommendation method for merging user social contact circle similarity according to claim 1, it is characterised in that:Institute State in S3 calculate user social contact circle similarity formula be<mrow> <msub> <mi>SSim</mi> <mrow> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> </mrow> </msub> <mo>=</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <munderover> <mo>&Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </munderover> <mo>&lsqb;</mo> <msub> <mi>p</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mo>&times;</mo> <mrow> <mo>(</mo> <msub> <mi>e</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>&rsqb;</mo> </mrow>Wherein pijRepresent user u1、u2The number of overlapping social circle's interior joint, eijRepresent user u1、u2It is overlapping in overlapping social circle The bar number on side;By user u1Social circle's recognition result be labeled as u1c1,u1c2,u1c3…u1cm, user u2Social circle identification knot Fruit is labeled as u2c1,u2c2,u2c3…u2cn, the degree of overlapping of one overlapping social circle of user is pij×(eij+ 1), m represents user u1 The social number of turns, n represents user u2The social number of turns.
- A kind of 6. friend recommendation method for merging user social contact circle similarity according to claim 1, it is characterised in that:Institute State the user property in S4, including the base attribute of user, the position of user, the online preference information of user.
- A kind of 7. friend recommendation method for merging user social contact circle similarity according to claim 6, it is characterised in that:Institute The base attribute for stating user is expressed asuserInfo(u1)={ basic (u1),tag(u1)}Wherein, basic (u1) represent user u1Personal information description, tag (u1) represent u1Label information;User u1, u2Base attribute similarity is designated assim(userInfo(u1),userInfo(u2))=λ1sim(basic(u1),basic(u2))+λ2sim(tag(u1),tag (u2))Wherein λ1+λ2=1, basic (u2) represent u2The personal information description of user, tag (u2) represent u2Label information.
- A kind of 8. friend recommendation method for merging user social contact circle similarity according to claim 6, it is characterised in that:Institute The information for stating the position of user is expressed as place (u1), user u1, u2Location similarity is sim (place (u1),place (u2));According to province, city, the similarities and differences for commonly using dwell point, sim (place (u are determined1),place(u2)) value.
- A kind of 9. friend recommendation method for merging user social contact circle similarity according to claim 6, it is characterised in that:Institute The online preference for stating user is expressed as net (u1), user u1, u2Preference similarity of surfing the Net is sim (net (u1),net(u2));Root User's cosine similarity often is calculated with app information, online duration, flow information according to user, represents that the online preference of user is similar Degree.
- 10. a kind of friend recommendation method of fusion user social contact circle similarity according to claim 6 or 7 or 8 or 9, its It is characterised by:The formula of calculating user's similarity is in the S4<mrow> <msub> <mi>USim</mi> <mrow> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> </mrow> </msub> <mo>=</mo> <msub> <mi>&lambda;</mi> <mn>1</mn> </msub> <mi>s</mi> <mi>i</mi> <mi>m</mi> <mrow> <mo>(</mo> <mi>b</mi> <mi>a</mi> <mi>s</mi> <mi>i</mi> <mi>c</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>)</mo> <mo>,</mo> <mi>b</mi> <mi>a</mi> <mi>s</mi> <mi>i</mi> <mi>c</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> <mo>)</mo> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>&lambda;</mi> <mn>2</mn> </msub> <mi>s</mi> <mi>i</mi> <mi>m</mi> <mrow> <mo>(</mo> <mi>p</mi> <mi>l</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>)</mo> <mo>,</mo> <mi>p</mi> <mi>l</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> <mo>)</mo> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>&lambda;</mi> <mn>3</mn> </msub> <mi>s</mi> <mi>i</mi> <mi>m</mi> <mrow> <mo>(</mo> <mi>n</mi> <mi>e</mi> <mi>t</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>1</mn> </msub> <mo>)</mo> <mo>,</mo> <mi>n</mi> <mi>e</mi> <mi>t</mi> <mo>(</mo> <msub> <mi>u</mi> <mn>2</mn> </msub> <mo>)</mo> <mo>)</mo> </mrow> </mrow>Relationship of Coefficients is λ1+λ2+λ3=1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710755695.8A CN107506455A (en) | 2017-08-29 | 2017-08-29 | A kind of friend recommendation method for merging user social contact circle similarity |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710755695.8A CN107506455A (en) | 2017-08-29 | 2017-08-29 | A kind of friend recommendation method for merging user social contact circle similarity |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107506455A true CN107506455A (en) | 2017-12-22 |
Family
ID=60692972
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710755695.8A Pending CN107506455A (en) | 2017-08-29 | 2017-08-29 | A kind of friend recommendation method for merging user social contact circle similarity |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107506455A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108509560A (en) * | 2018-03-23 | 2018-09-07 | 广州杰赛科技股份有限公司 | User's similarity preparation method and device, equipment, storage medium |
CN109766940A (en) * | 2018-12-29 | 2019-05-17 | 北京天诚同创电气有限公司 | The method and apparatus for assessing the similarity between multiple sewage disposal systems |
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 |
CN110489664A (en) * | 2019-08-07 | 2019-11-22 | 北京达佳互联信息技术有限公司 | A kind of client recommended method and device based on compatible degree |
CN113269653A (en) * | 2021-06-18 | 2021-08-17 | 北京市科学技术情报研究所 | Social network management method and system based on circled thought |
CN115357632A (en) * | 2022-07-21 | 2022-11-18 | 广东技术师范大学 | Learner personalized interpretable potential friend recommendation method and device |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103116614A (en) * | 2013-01-25 | 2013-05-22 | 北京奇艺世纪科技有限公司 | Collaborative filtering recommendation method, device and system base on user track |
CN103995823A (en) * | 2014-03-25 | 2014-08-20 | 南京邮电大学 | Information recommending method based on social network |
CN104731962A (en) * | 2015-04-03 | 2015-06-24 | 重庆邮电大学 | Method and system for friend recommendation based on similar associations in social network |
US20150281163A1 (en) * | 2014-03-27 | 2015-10-01 | International Business Machines Corporation | User recommendations in a social media network |
CN106339483A (en) * | 2016-08-30 | 2017-01-18 | 电子科技大学 | Social activity recommendation algorithm in mobile social network |
-
2017
- 2017-08-29 CN CN201710755695.8A patent/CN107506455A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103116614A (en) * | 2013-01-25 | 2013-05-22 | 北京奇艺世纪科技有限公司 | Collaborative filtering recommendation method, device and system base on user track |
CN103995823A (en) * | 2014-03-25 | 2014-08-20 | 南京邮电大学 | Information recommending method based on social network |
US20150281163A1 (en) * | 2014-03-27 | 2015-10-01 | International Business Machines Corporation | User recommendations in a social media network |
CN104731962A (en) * | 2015-04-03 | 2015-06-24 | 重庆邮电大学 | Method and system for friend recommendation based on similar associations in social network |
CN106339483A (en) * | 2016-08-30 | 2017-01-18 | 电子科技大学 | Social activity recommendation algorithm in mobile social network |
Non-Patent Citations (3)
Title |
---|
VINCENT D. BLONDEL等: "Fast unfolding of communities in large networks", 《JOURNAL OF STATISTICAL MECHANICS: THEORY AND EXPERIMENT》 * |
徐志明 等: "微博用户的相似性度量及其应用", 《计算机学报》 * |
王玙 等: "基于社交圈的在线社交网络朋友推荐算法", 《计算机学报》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN108509560A (en) * | 2018-03-23 | 2018-09-07 | 广州杰赛科技股份有限公司 | User's similarity preparation method and device, equipment, storage medium |
CN108509560B (en) * | 2018-03-23 | 2021-04-09 | 广州杰赛科技股份有限公司 | User similarity obtaining method and device, equipment and storage medium |
CN109766940A (en) * | 2018-12-29 | 2019-05-17 | 北京天诚同创电气有限公司 | The method and apparatus for assessing the similarity between multiple sewage disposal systems |
CN109766940B (en) * | 2018-12-29 | 2024-02-02 | 北京天诚同创电气有限公司 | Method and apparatus for evaluating similarity between multiple sewage treatment systems |
CN110489664A (en) * | 2019-08-07 | 2019-11-22 | 北京达佳互联信息技术有限公司 | A kind of client recommended method and device based on compatible degree |
CN113269653A (en) * | 2021-06-18 | 2021-08-17 | 北京市科学技术情报研究所 | Social network management method and system based on circled thought |
CN113269653B (en) * | 2021-06-18 | 2024-03-29 | 北京市科学技术情报研究所 | Social network management method and system based on layering thought |
CN115357632A (en) * | 2022-07-21 | 2022-11-18 | 广东技术师范大学 | Learner personalized interpretable potential friend recommendation method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107506455A (en) | A kind of friend recommendation method for merging user social contact circle similarity | |
CN103514204B (en) | Information recommendation method and device | |
CN103020116B (en) | The method of the powerful user of automatic screening on social media network | |
CN109166017A (en) | Method for pushing, device, computer equipment and storage medium based on reunion class | |
CN103353872B (en) | A kind of teaching resource personalized recommendation method based on neutral net | |
CN112115377B (en) | Graph neural network link prediction recommendation method based on social relationship | |
CN110263256A (en) | Personalized recommendation method based on multi-modal heterogeneous information | |
CN105843799B (en) | A kind of academic paper label recommendation method based on multi-source heterogeneous information graph model | |
CN103823888B (en) | Node-closeness-based social network site friend recommendation method | |
CN104077417B (en) | People tag in social networks recommends method and system | |
CN108665323A (en) | A kind of integrated approach for finance product commending system | |
CN107944035A (en) | A kind of image recommendation method for merging visual signature and user's scoring | |
CN109190030A (en) | Implicit feedback recommendation method fusing node2vec and deep neural network | |
CN103942308A (en) | Method and device for detecting large-scale social network communities | |
CN105354260B (en) | The Mobile solution of a kind of mosaic society's network and item characteristic recommends method | |
CN106934071A (en) | Recommendation method and device based on Heterogeneous Information network and Bayes's personalized ordering | |
CN104572733B (en) | The method and device of user interest labeling | |
CN108647216A (en) | Software crowdsourcing task recommendation system and method based on developer's social networks | |
CN111949887A (en) | Item recommendation method and device and computer-readable storage medium | |
CN107545471A (en) | A kind of big data intelligent recommendation method based on Gaussian Mixture | |
CN107507054A (en) | A kind of proposed algorithm based on Recognition with Recurrent Neural Network | |
CN111104604A (en) | Lightweight social recommendation method based on Hash learning | |
CN108876536A (en) | Collaborative filtering recommending method based on arest neighbors information | |
CN105869058B (en) | A kind of method that multilayer latent variable model user portrait extracts | |
CN102129457A (en) | Method for inquiring large-scale semantic data paths |
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: 20171222 |
|
RJ01 | Rejection of invention patent application after publication |