CN104092567B - Determine the method and apparatus of the influence power sequence of user - Google Patents
Determine the method and apparatus of the influence power sequence of user Download PDFInfo
- Publication number
- CN104092567B CN104092567B CN201410294986.8A CN201410294986A CN104092567B CN 104092567 B CN104092567 B CN 104092567B CN 201410294986 A CN201410294986 A CN 201410294986A CN 104092567 B CN104092567 B CN 104092567B
- Authority
- CN
- China
- Prior art keywords
- user
- message
- target topic
- information
- ranking value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 239000013598 vector Substances 0.000 claims abstract description 97
- 239000011159 matrix material Substances 0.000 claims abstract description 75
- 230000007704 transition Effects 0.000 claims abstract description 70
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 19
- 239000004744 fabric Substances 0.000 claims 2
- 238000004364 calculation method Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 9
- 238000003860 storage Methods 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- JTJMJGYZQZDUJJ-UHFFFAOYSA-N phencyclidine Chemical compound C1CCCCN1C1(C=2C=CC=CC=2)CCCCC1 JTJMJGYZQZDUJJ-UHFFFAOYSA-N 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Landscapes
- Information Transfer Between Computers (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment provides a kind of method and apparatus for the influence power for determining user.This method includes：The first information and the second information are obtained, the first information is used for the theme of the mutual concern relation for indicating N number of user in social networks, message count of second information for indicating each user's issue in N number of user, and the message of each user issue；According to the first information and the second information, the corresponding N-dimensional weight vectors of target topic are determined；The 3rd information is obtained, the 3rd information is used to indicate to forward the quantity of message and the theme of forwarding message between the user that there is message to forward relation in N number of user；According to the 3rd information, the corresponding transition probability matrix of target topic is determined；According to transition probability matrix and weight vector, determined using Page sort algorithm in social networks, sequence of N number of user in the influence power in target topic field.Due to reflecting the degree of correlation actual between user in social networks, therefore it is effectively improved the accuracy of influence power ranking results.
Description
Technical field
The present invention relates to computer network field, and more particularly to it is a kind of determine user influence power sequence method and dress
Put.
Background technology
With the popularization of fast wireless network, the network media is propagated in social information and commercial field plays and increasingly weighed
The role wanted.Progressively substitution traditional media turns into the most important of Information Communication for such as integrated information portal website and social networks
The larger website of influence power and user are even more to play promotion in the propagation of event in platform, these network informations and social platform
Effect.
The network node of material impact (for example can be being found from intricate and huge user relational network
Website, cellphone subscriber, microblogging or wechat user, carrier gateway etc.) it is the key that product promotion or Information Communication are monitored.Therefore
It is one kind that enterprise carries out product promotion or Information Communication monitoring to find out influential network node and carry out influence power ranking
Effective method.
PageRank (Page sequence) algorithm is the general-purpose algorithm of network node influence power ranking, and it is carried by Google at first
Go out.Page sort algorithm defines random surfer model, that is, imagine network in there are many random surfers, they from certain
Individual network node (for example, webpage) is set out, the migration between network node according to transition probability matrix, then the position of random surfer
Put the importance that probability distribution represents webpage.
Page sort algorithm has the Page sequence (Topic Sensitive PageRank) of prototype version and subject-oriented
Version.Wherein, the Page sort algorithm of subject-oriented can be applied in social networks, for determining that each user exists in social networks
The influence power of a certain target topic.The algorithm obtains the influence power sequence of each user using equation below iteration：
pi+1=β Mpi+(1-β)es/│S│
Wherein, pi+1And piWhen being illustrated respectively in i+1 time and ith iteration calculating, the Page ranking value composition of each user
Vector, the vectorial dimension n represents the number of user to be sorted；β is the transfer constant between 0 to 1, usual value
0.15；M is transition probability matrix, is set up based on the mutual concern relation between user；E in formulas/ │ S │ be weight to
Amount, wherein, S is the set of the related node of target topic, esThe value of each element depends on the corresponding user of the element in vector
Whether set S is belonged to, if it is, taking 1, otherwise, take 0.
But, in social networks, the Page sort algorithm of above-mentioned subject-oriented has the following disadvantages：The choosing of weight vectors
Select it is simple use 0 or 1, i.e., related or uncorrelated expression, and in practice, there is also correlation between the user related to target topic
The difference of degree, the expression of above-mentioned weight vector is not accurate enough.In addition, each element is only by between user in transition probability matrix
Concern relation determine.It can not reflect in actual social networks, so that the accuracy of influence power ranking results is relatively low.
The content of the invention
The embodiment provides a kind of method and apparatus for the influence power for determining user, shadow can be effectively improved
Ring the accuracy of power ranking results.
In a first aspect, providing a method that, this method includes：The first information and the second information are obtained, the first information is used
In the mutual concern relation for indicating N number of user in social networks, the second information is used to indicate each user's issue in N number of user
Message count, and each user issue message theme；According to the first information and the second information, the corresponding N of target topic is determined
The i-th element in dimensional weight vector, weight vectors is determined based on the first proportion and the second proportion, wherein, the first proportion is：
In the message of i-th user issue, on the proportion shared by the message of target topic, the second proportion is：The user of i-th user concern
In, the message package of issue contains on the proportion shared by the user of the message of target topic, and i is 1 arbitrary integer into N；Obtain
3rd information, the 3rd information be used for indicate in N number of user have message forward relation user between forwarding message quantity and
Forward the theme of message；According to the 3rd information, the corresponding transition probability matrix of target topic, the of transition probability matrix are determined
(j, k) element is to be based on：Kth user is forwarded by jth user, and on the quantity of the message of target topic, jth user is corresponding
The size between the quantity of association user, and the quantity on the message of target topic of forwarding association user issue is forwarded to close
What system determined, wherein, forwarding association user is in N number of user, the message of issue is by the forwarded over user of jth user, and j, k are
1 arbitrary integer into N；According to transition probability matrix and weight vector, determined using Page sort algorithm in social networks,
Sequence of N number of user in the influence power in target topic field.
With reference in a first aspect, in the first possible implementation of first aspect, the i-th element in weight vectors is
Determined by below equation：Ai=x (a/b)+(1-x) (c/d), wherein, AiIt is the i-th element in weight vectors；X is advance
The real number of setting, and 0≤x≤1；A is the quantity for the message on target topic that the i-th user issues；B issues for the i-th user
Message sum；In the user that c pays close attention to for the i-th user, the message package of issue contains the number of the user of the message on target topic
Amount；D is the total number of users that the i-th user pays close attention to.
With reference to the first possible implementation of first aspect or first aspect, second in first aspect is possible
In implementation, (j, k) element of transition probability matrix is determined by below equation：Bjk=max (e, 1)/max (f,
G), wherein, BjkFor (j, k) element of transition probability matrix；E is what kth user was forwarded by jth user, on target topic
Message quantity；The quantity for the message on target topic that f issues for forwarding association user；G is corresponding turn of jth user
Send out the quantity of association user.
With reference to the first or second of possible implementation of first aspect, first aspect, the 3rd of first aspect the
Plant in possible implementation, according to transition probability matrix and weight vector, the shadow of N number of user is determined using Page sort algorithm
The sequence of power is rung, including：According to transition probability matrix and weight vector, self-corresponding Page ranking value each to N number of user is carried out
First time iteration, first time iteration is iterations limited number of time iteration set in advance；
After first time iteration, N number of each self-corresponding Page ranking value of user, by N number of user be divided into first set and
Second set, wherein, the corresponding Page ranking value of user in second set is less than the corresponding Page of user in first set
Ranking value；According to transition probability matrix and weight vector, self-corresponding Page ranking value each to N number of user carries out second repeatedly
In generation, in second of iteration, the iteration stopping threshold value of the corresponding Page ranking value of user in first set is less than second set
In the corresponding Page ranking value of user iteration stopping threshold value；After second of iteration, N number of each self-corresponding Page row of user
The size sequence of sequence value is defined as sequence of N number of user in the influence power in target topic field.
Second aspect includes there is provided a kind of method, this method：Self-corresponding Page ranking value each to N number of network node
First time iteration is carried out, first time iteration is iterations limited number of time iteration set in advance；It is N number of after first time iteration
Each self-corresponding Page ranking value of network node, is divided into first set and second set by N number of network node, wherein, the second collection
The corresponding Page ranking value of network node in conjunction is less than the corresponding Page ranking value of network node in first set；To N number of
Each self-corresponding Page ranking value of network node carries out second of iteration, in second of iteration, the network section in first set
The iteration stopping threshold value of the corresponding Page ranking value of point is less than changing for the corresponding Page ranking value of network node in second set
For outage threshold；After second of iteration, the sequence of the size of N number of each self-corresponding Page ranking value of network node is defined as N number of
The sequence of the influence power of network node.
The third aspect includes there is provided a kind of device, the device：First acquisition module, for obtaining the first information and
Two information, the first information is used for the mutual concern relation for indicating N number of user in social networks, and the second information is N number of for indicating
The message count of each user issue in user, and the message of each user issue theme；First determining module, for according to first
The first information and the second information that acquisition module is obtained, determine the in the corresponding N-dimensional weight vectors of target topic, weight vectors
I elements are determined based on the first proportion and the second proportion, wherein, the first proportion is：In the message of i-th user issue, on
Proportion shared by the message of target topic, the second proportion is：In the user of i-th user concern, the message package of issue contains on mesh
Proportion shared by the user for the message for marking theme, i is 1 arbitrary integer into N；Second acquisition module, believes for obtaining the 3rd
Breath, the 3rd information is used to indicate to forward the quantity of message and forwarding to disappear between the user that there is message to forward relation in N number of user
The theme of breath；Second determining module, for the 3rd information obtained according to the second acquisition module, determines corresponding turn of target topic
Probability matrix is moved, (j, k) element of transition probability matrix is to be based on：Kth user is forwarded by jth user, on target master
The quantity of the message of topic, jth user it is corresponding forwarding association user quantity, and forwarding association user issue on target
What magnitude relationship between the quantity of the message of theme was determined, wherein, during forwarding association user is N number of user, the message of issue
By the forwarded over user of jth user, j, k are 1 arbitrary integer into N；3rd determining module, for according to transition probability square
Battle array and weight vector, are determined in social networks using Page sort algorithm, influence power of N number of user in target topic field
Sequence.
With reference to the third aspect, in the first possible implementation of the third aspect, the i-th element in weight vectors is
Determined by below equation：Ai=x (a/b)+(1-x) (c/d), wherein, AiIt is the i-th element in weight vectors；X is advance
The real number of setting, and 0≤x≤1；A is the quantity for the message on target topic that the i-th user issues；B issues for the i-th user
Message sum；In the user that c pays close attention to for the i-th user, the message package of issue contains the number of the user of the message on target topic
Amount；D is the total number of users that the i-th user pays close attention to.
With reference to the first possible implementation of the third aspect or the third aspect, second in the third aspect is possible
In implementation, (j, k) element of transition probability matrix is determined by below equation：Bjk=max (e, 1)/max (f,
G), wherein, BjkFor (j, k) element of transition probability matrix；E is what kth user was forwarded by jth user, on target topic
Message quantity；The quantity for the message on target topic that f issues for forwarding association user；G is corresponding turn of jth user
Send out the quantity of association user.
With reference to the first or second of possible implementation of the third aspect, the third aspect, the 3rd of the third aspect the
Plant in possible implementation, the 3rd determining unit is each to N number of user specifically for according to transition probability matrix and weight vector
Self-corresponding Page ranking value carries out first time iteration, and first time iteration is iterations limited number of time iteration set in advance；Root
After first time iteration, N number of user is divided into first set and second set by N number of each self-corresponding Page ranking value of user, its
In, the corresponding Page ranking value of user in second set is less than the iteration of the corresponding Page ranking value of user in first set
Outage threshold；According to transition probability matrix and weight vector, self-corresponding Page ranking value each to N number of user carries out second repeatedly
In generation, in second of iteration, the iteration stopping threshold value of the corresponding Page ranking value of user in first set is less than second set
In the corresponding Page ranking value of user iteration stopping threshold value；After second of iteration, N number of each self-corresponding Page row of user
The size sequence of sequence value is defined as sequence of N number of user in the influence power in target topic field.
Fourth aspect includes there is provided a kind of device, the device：First iteration module, is each corresponded to N number of network node
Page ranking value carry out first time iteration, first time iteration be iterations limited number of time iteration set in advance；Determining module,
Each self-corresponding Page ranking value of N number of network node for being obtained according to the first iteration module, is divided into by N number of network node
One set and second set, wherein, the corresponding Page ranking value of network node in second set is less than the net in first set
The corresponding Page ranking value of network node；Secondary iteration module, N number of network node for being obtained to the first iteration module is each right
The Page ranking value answered carries out second of iteration, in second of iteration, the network section in the first set that determining module is determined
The iteration stopping threshold value of the corresponding Page ranking value of point is less than changing for the corresponding Page ranking value of network node in second set
For outage threshold；Determining module is additionally operable to after second of iteration, the size of N number of each self-corresponding Page ranking value of network node
Sequence is defined as the sequence of the influence power of N number of network node.
Embodiments of the invention are employed (for example forwards relation, concern relation based on social networks actual between user
And target topic) transition probability matrix that determines and weight vectors carry out influence power sequence.Due to the reality between user
Social networks reflect the degree of correlation actual between user in social networks, therefore be effectively improved influence power ranking results
Accuracy.
Brief description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, will make below to required in the embodiment of the present invention
Accompanying drawing is briefly described, it should be apparent that, drawings described below is only some embodiments of the present invention, for
For those of ordinary skill in the art, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings
Accompanying drawing.
Fig. 1 is the indicative flowchart of the method for the influence power of determination user according to embodiments of the present invention.
Fig. 2 is the schematic diagram of the device of the influence power of determination user according to another embodiment of the present invention.
Fig. 3 is the indicative flowchart of the method for the influence power of determination user according to another embodiment of the present invention.
Fig. 4 is the schematic diagram of the device of the influence power of the determination user according to further embodiment of this invention.
Fig. 5 is the indicative flowchart of the method for the influence power of determination network node according to another embodiment of the present invention.
Fig. 6 is the schematic diagram of the device of the influence power of determination network node according to another embodiment of the present invention.
Fig. 7 is the indicative flowchart of the method for the influence power of determination network node according to another embodiment of the present invention.
Fig. 8 is the schematic diagram of the device of the influence power of determination network node according to another embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is a part of embodiment of the present invention, rather than whole embodiments.Based on this hair
Embodiment in bright, the every other reality that those of ordinary skill in the art are obtained on the premise of creative work is not made
Example is applied, should all belong to the scope of protection of the invention.
Fig. 1 is the indicative flowchart of the method for the influence power of determination user according to embodiments of the present invention.Fig. 1 method
Performed by the device for the influence power for determining user.Fig. 1 method includes：
110th, the first information and the second information are obtained, the first information is used to indicate that N number of user's in social networks to be mutual
The master of concern relation, message count of second information for indicating each user's issue in N number of user, and the message of each user issue
Topic；
120th, according to the first information and the second information, determine in the corresponding N-dimensional weight vectors of target topic, weight vectors
I-th element is determined based on the first proportion and the second proportion, wherein, the first proportion is：In the message of i-th user issue, close
Proportion shared by message in target topic, the second proportion is：In the user of i-th user concern, the message package of issue contain on
Proportion shared by the user of the message of target topic, i is 1 arbitrary integer into N；
130th, the 3rd information is obtained, the 3rd information is used to indicate between the user with message forwarding relation in N number of user
Forward the quantity of message and the theme of forwarding message；
140th, according to the 3rd information, the corresponding transition probability matrix of target topic, (j, k) of transition probability matrix are determined
Element is to be based on：Kth user is forwarded by jth user, on the quantity of the message of target topic, the corresponding forwarding of jth user
Magnitude relationship between the quantity of association user, and the quantity on the message of target topic of forwarding association user issue is true
Fixed, wherein, forwarding association user is in N number of user, the message of issue is by the forwarded over user of jth user, and j, k are 1 to N
In arbitrary integer；
150th, according to transition probability matrix and weight vector, determined using Page sort algorithm in social networks, it is N number of to use
Sequence of the family in the influence power in target topic field.
It should be understood that the second information also can indicate that the message sum of each user issue.The master of the message of each user's issue
Topic can be such as society, education, economic, military, amusement in one or more.
Embodiments of the invention are employed (for example forwards relation, concern relation based on social networks actual between user
And target topic) transition probability matrix that determines and weight vectors carry out influence power sequence.Due to the reality between user
Social networks reflect the degree of correlation actual between user in social networks, therefore be effectively improved influence power ranking results
Accuracy.
The i-th element in embodiments in accordance with the present invention, weight vectors is determined by below equation：Ai=x (a/b)
+ (1-x) (c/d), wherein, AiIt is the i-th element in weight vectors；X is real number set in advance, and 0≤x≤1；A is the i-th use
The quantity of the message on target topic of family issue；B is the message sum that the i-th user issues；C is the use that the i-th user pays close attention to
In family, the message package of issue contains the quantity of the user of the message on target topic；D is the total number of users that the i-th user pays close attention to.
It should be understood that determining that the element in weight vectors can also use the other formula determined based on a/b and c/d.Example
Such as simply deformed.
Embodiments in accordance with the present invention, (j, k) element of transition probability matrix is determined by below equation：Bjk=
Max (e, 1)/max (f, g), wherein, BjkFor (j, k) element of transition probability matrix；E is that kth user is forwarded by jth user
, the quantity on the message of target topic；The quantity for the message on target topic that f issues for forwarding association user；G is
The quantity of the corresponding forwarding association user of jth user.
It should be understood that determining that the element in transition probability matrix can also use the other formula determined based on e, f and g.
Embodiments in accordance with the present invention, according to transition probability matrix and weight vector, are determined N number of using Page sort algorithm
The sequence of the influence power of user, including：According to transition probability matrix and weight vector, self-corresponding Page row each to N number of user
Sequence value carries out first time iteration, and first time iteration is iterations limited number of time iteration set in advance；After first time iteration,
N number of each self-corresponding Page ranking value of user, is divided into first set and second set by N number of user, wherein, in second set
The corresponding Page ranking value of user is less than the corresponding Page ranking value of user in first set；According to transition probability matrix and power
Value vector, self-corresponding Page ranking value each to N number of user carries out second of iteration, in second of iteration, in first set
The corresponding Page ranking value of user iteration stopping threshold value be less than second set in the corresponding Page ranking value of user repeatedly
For outage threshold；After second of iteration, the size sequence of N number of each self-corresponding Page ranking value of user is defined as N number of user
In the sequence of the influence power in target topic field.
It should be understood that iterations limited number of time iteration set in advance, can be the iteration according to Six Degrees theory setting
Number of times, such as 5 times.Can also be other iterationses, such as less than 10 times.In other words, first time iteration is iteration time
Number limited number of time iteration set in advance, the preliminary row of influence power is carried out in order to the ranking value of the iteration result according to first time
Sequence, and the user of first set is subjected to more accurate sequence as needed.Fig. 1 method can also include：For the first time
In iteration, to the Page ranking value setting convergence outage threshold of N number of user.After first time iteration, N number of user each corresponds to
Page ranking value, N number of user is divided into first set and second set, Ke Yishi determines two according to power-law distribution rule
In Page ranking value after set, such as first time iteration, according to size order, preceding 20% user is first set, after
80% user is second set.Different set can also be divided as needed.
Because the convergence outage threshold set to first set is less than the convergence outage threshold that second set is set, therefore
When being sorted to influence power, the computing of less time is carried out to second set, so as to reduce the expense of calculating.
In other words, social network data has worldlet and big data attribute, and the influence power ranking value of user node is in
Power-law distribution, i.e., the influence force value of most nodes is all very low, and the influence power of only small part node is very high.In actual section
In point influence power ranking application, and all nodes need not be carried out with ranking, only need to the node big to a small amount of influence power arrange
Name.The small node of most of influence power is after recursive calculation 5 times, it is possible to ignore the change of these nodes, it is no longer necessary to calculate,
So as to largely save the calculating time.
The method that described above is the influence power of determination user according to embodiments of the present invention, describes root below in conjunction with the accompanying drawings
According to the device of the influence power of the determination user of the embodiment of the present invention.
Fig. 2 is the schematic diagram of the device of the influence power of determination user according to another embodiment of the present invention.Fig. 2's
Device 200 includes：First acquisition module 210, for obtaining the first information and the second information, the first information is used to indicate social network
The mutual concern relation of N number of user in network, the second information is used to indicate the message count of each user's issue in N number of user, and
The theme of the message of each user's issue；First determining module 220, for the first information obtained according to the first acquisition module 210
With the second information, determine the i-th element in the corresponding N-dimensional weight vectors of target topic, weight vectors be based on the first proportion and
What the second proportion was determined, wherein, the first proportion is：In the message of i-th user issue, on the ratio shared by the message of target topic
Heavy, the second proportion is：In the user of i-th user concern, the message package of issue contains on shared by the user of the message of target topic
Proportion, i be 1 arbitrary integer into N；Second acquisition module 230, for obtaining the 3rd information, the 3rd information is used to indicate N
There is message to forward the quantity of message and the theme of forwarding message between forwarding the user of relation in individual user；Second determining module
240, for the 3rd information obtained according to the second acquisition module 230, the corresponding transition probability matrix of target topic is determined, is shifted
(j, k) element of probability matrix is to be based on：Kth user is forwarded by jth user, on the quantity of the message of target topic,
The quantity of the corresponding forwarding association user of jth user, and forward the number of the message on target topic of association user issue
What the magnitude relationship between amount was determined, wherein, forwarding association user is in N number of user, the message of issue is forwarded over by jth user
User, j, k are 1 arbitrary integer into N；3rd determining module 250, for according to transition probability matrix and weights to
Amount, is determined in social networks, sequence of N number of user in the influence power in target topic field using Page sort algorithm.
Embodiments of the invention are employed (for example forwards relation, concern relation based on social networks actual between user
And target topic) transition probability matrix that determines and weight vectors carry out influence power sequence.Due to the reality between user
Social networks reflect the degree of correlation actual between user in social networks, therefore be effectively improved influence power ranking results
Accuracy.
The i-th element in embodiments in accordance with the present invention, weight vectors is determined by below equation：Ai=x (a/b)
+ (1-x) (c/d), wherein, AiIt is the i-th element in weight vectors；X is real number set in advance, and 0≤x≤1；A is the i-th use
The quantity of the message on target topic of family issue；B is the message sum that the i-th user issues；C is the use that the i-th user pays close attention to
In family, the message package of issue contains the quantity of the user of the message on target topic；D is the total number of users that the i-th user pays close attention to.
Embodiments in accordance with the present invention, (j, k) element of transition probability matrix is determined by below equation：Bjk=
Max (e, 1)/max (f, g), wherein, BjkFor (j, k) element of transition probability matrix；E is that kth user is forwarded by jth user
, the quantity on the message of target topic；The quantity for the message on target topic that f issues for forwarding association user；G is
The quantity of the corresponding forwarding association user of jth user.
Embodiments in accordance with the present invention, the 3rd determining unit is specifically for according to transition probability matrix and weight vector, to N
Individual each self-corresponding Page ranking value of user carries out first time iteration, and first time iteration is iterations limited number of time set in advance
Iteration；After first time iteration, N number of user is divided into first set and second by N number of each self-corresponding Page ranking value of user
Set, wherein, the corresponding Page ranking value of user in second set is less than the corresponding Page sequence of user in first set
The iteration stopping threshold value of value；According to transition probability matrix and weight vector, self-corresponding Page ranking value each to N number of user is carried out
Second of iteration, in second of iteration, the iteration stopping threshold value of the corresponding Page ranking value of user in first set is less than
The iteration stopping threshold value of the corresponding Page ranking value of user in second set；After second of iteration, N number of user each corresponds to
Page ranking value size sequence be defined as sequence of N number of user in the influence power in target topic field.
The method that the operation and function for determining the modules of the device 200 of the influence power of user may be referred to above-mentioned Fig. 1,
In order to avoid repeating, it will not be repeated here.With reference to specific example, embodiments of the invention are described more fully.
Fig. 3 is the indicative flowchart of the method for the influence power of determination user according to another embodiment of the present invention.Fig. 3's
Method is the example of Fig. 1 method.In second set in the embodiment of user's corresponding diagram 1 in set Z in the present embodiment
User, the user in the first set in the embodiment of user's corresponding diagram 1 in set Z is not belonging in set Y；First threshold
Use in the iteration stopping threshold value of the corresponding Page ranking value of user in correspondence first set, Second Threshold correspondence first set
The iteration stopping threshold value of the corresponding Page ranking value in family.First ranking value is used for the initialization of ranking value vector, wherein each use
First ranking value at family is the element of ranking value vector；N number of user is each right after second ranking value vector correspondence first time iteration
The vector that the Page ranking value answered is constituted.Concrete implementation mode is as follows.
310th, iterations is set.
For example it is theoretical based on Six Degrees, iterations is set as 5, in other words, node says the transmission of influence power most
It is many just to reach all nodes no more than 5 times.
320th, to the Page ranking value assignment of all users.
Specifically, the ranking value of identical first is set to all users, for example, there will be user to be entered as 1 more.
330th, first threshold is set.
The first threshold is used for the fluctuation ratio with the influence power ranking value of user compared with the fluctuation of influence power ranking value is defined as
R=│ pk-pk+1│/pk。
340th, calculated according to the iterations of setting, and determine that the influence power of certain customers sorts according to first threshold
Value.
It is for instance possible to use formula 3.1:pk+1=β Bpk+(1-β)A (3.1)
Wherein, B is the transition probability matrix of social class theme, and the matrix is N N matrix；A is the weight of social class theme
Vector, the vector is vectorial for the N-dimensional of the N number of user of correspondence；β is constant, for example, can be 0.15；pkAnd pk+1Respectively user
Ranking value vector of the ranking value after kth time and+1 iteration of kth, the vector is the N-dimensional vector of the N number of user of correspondence.Weight
The i-th element in vectorial A is determined by below equation：Ai=x (a/b)+(1-x) (c/d), wherein, AiIn being weight vectors
The i-th element；X is real number set in advance, and 0≤x≤1；A is the number for the message on target topic that the i-th user issues
Amount；B is the message sum that the i-th user issues；In the user that c pays close attention to for the i-th user, the message package of issue contains on target topic
Message user quantity；D is the total number of users that the i-th user pays close attention to.Transition probability matrix B (j, k) element is to pass through
What below equation was determined：Bjk=max (e, 1)/max (f, g), wherein, BjkFor (j, k) element of transition probability matrix；E is
Kth user is forwarded by jth user, the quantity on the message of target topic；F is forwarding association user issue on target
The quantity of the message of theme；G is the quantity of the corresponding forwarding association user of jth user.
Specifically, the first ranking value vector of user is carried out using the weight vectors A of transition probability matrix B and theme
Iterative calculation, and calculate after each iterative calculation the undulating value R of the ranking value of each user.By undulating value R and the first threshold
Value compares：If the undulating value in 5 iterative calculation is not more than first threshold, stop the calculating to the user.Simultaneously will
The ranking value of the user is defined as influence power ranking value, and the user is defined as into the element in set X.If in 5 iteration
After calculating, the undulating value of the ranking value of user is more than first threshold, then the user is defined as into element in set Y, and perform
350。
350th, the second ranking value vector of user is determined.
Specifically, 5 iterative calculation are carried out by multiple first ranking values vector to transition probability matrix and user, really
The second ranking value vector of user is determined, wherein the element in the second ranking value vector is the second ranking value of each user.
360th, the set Z of low influence power user in set Y is determined.
Specifically, tentatively sorted, and tied according to sequence according to the size of the second ranking value of each user in set Y
Fruit determines the set Z of low influence power user.For example, sequence is defined as into low influence power in rear 80% user according to power-law distribution
User.
Similar to 340 for the user being not belonging in set Y in set Z, being determined according to first threshold and formula 3.1 should
The influence power ranking value of user.370 are performed to the user in set Z.
370th, Second Threshold is set for the user in set Z.
Specifically, Second Threshold of the setting more than first threshold.Due to sort rear 80% user be low influence power use
Family, therefore Second Threshold is not reduced the accuracy of final influence power sequence substantially more than first threshold.
380th, the influence power ranking value of user in set Z is determined according to Second Threshold.
For example, according to formula 3.1, using the weight vectors A of transition probability matrix B and theme to the second ranking value of user
Vector is iterated calculating, and each undulating value of the ranking value of user in set of computations Z after iterating to calculate every time.By the ripple
Dynamic value is compared with Second Threshold：If the undulating value is not more than Second Threshold, stop the calculating to the user.This is used simultaneously
The ranking value at family is defined as the influence power ranking value of user in set Z.
390th, the influence power of all users is determined according to the influence power ranking value of all users.
Specifically, can be by user according to influence power ranking value because the influence power of the big user of influence power ranking value is high
Sort to determine the influence power of all users from big to small.
Fig. 4 is the indicative flowchart of the device of the influence power of determination user according to another embodiment of the present invention.Fig. 4's
Device 400 includes：Processor 410, memory 420, communication bus 430.Processor 410 calls storage by communication bus 430
Code in device 420, and be used for：The first information and the second information are obtained, the first information is used to indicate N number of use in social networks
The mutual concern relation at family, message count of second information for indicating each user's issue in N number of user, and each user issue
The theme of message；According to the first information and the second information, determine in the corresponding N-dimensional weight vectors of target topic, weight vectors
I-th element is determined based on the first proportion and the second proportion, wherein, the first proportion is：In the message of i-th user issue, close
Proportion shared by message in target topic, the second proportion is：In the user of i-th user concern, the message package of issue contain on
Proportion shared by the user of the message of target topic, i is 1 arbitrary integer into N；The 3rd information is obtained, the 3rd information is used for
Indicate to forward the quantity of message and the theme of forwarding message between the user that there is message to forward relation in N number of user；According to
Three information, determine the corresponding transition probability matrix of target topic, (j, k) element of transition probability matrix is to be based on：Kth user
Forwarded by jth user, on the quantity of the message of target topic, the quantity of the corresponding forwarding association user of jth user, and
Forward what the magnitude relationship between the quantity on the message of target topic of association user issue was determined, wherein, forwarding association
User is in N number of user, the message of issue is by the forwarded over user of jth user, and j, k are 1 arbitrary integer into N；According to
Transition probability matrix and weight vector, determine that in social networks N number of user is in target topic field using Page sort algorithm
Influence power sequence.
Embodiments of the invention are employed (for example forwards relation, concern relation based on social networks actual between user
And target topic) transition probability matrix that determines and weight vectors carry out influence power sequence.Due to the reality between user
Social networks reflect the degree of correlation actual between user in social networks, therefore be effectively improved influence power ranking results
Accuracy.
The i-th element in embodiments in accordance with the present invention, weight vectors is determined by below equation：Ai=x (a/b)
+ (1-x) (c/d), wherein, AiIt is the i-th element in weight vectors；X is real number set in advance, and 0≤x≤1；A is the i-th use
The quantity of the message on target topic of family issue；B is the message sum that the i-th user issues；C is the use that the i-th user pays close attention to
In family, the message package of issue contains the quantity of the user of the message on target topic；D is the total number of users that the i-th user pays close attention to.
Embodiments in accordance with the present invention, (j, k) element of transition probability matrix is determined by below equation：Bjk=
Max (e, 1)/max (f, g), wherein, BjkFor (j, k) element of transition probability matrix；E is that kth user is forwarded by jth user
, the quantity on the message of target topic；The quantity for the message on target topic that f issues for forwarding association user；G is
The quantity of the corresponding forwarding association user of jth user.
Embodiments in accordance with the present invention, processor 410 specifically for：According to transition probability matrix and weight vector, to N number of
Each self-corresponding Page ranking value of user carries out first time iteration, and first time iteration is that iterations limited number of time set in advance changes
Generation；After first time iteration, N number of user is divided into first set and the second collection by N number of each self-corresponding Page ranking value of user
Close, wherein, the corresponding Page ranking value of user in second set is less than the corresponding Page ranking value of user in first set；
According to transition probability matrix and weight vector, self-corresponding Page ranking value each to N number of user carries out second of iteration, second
In secondary iteration, the iteration stopping threshold value of the corresponding Page ranking value of user in first set is less than the user couple in second set
The iteration stopping threshold value for the Page ranking value answered；After second of iteration, the size of N number of each self-corresponding Page ranking value of user
Sequence is defined as sequence of N number of user in the influence power in target topic field.
The method that the operation and function for determining the modules of the device 400 of the influence power of user may be referred to above-mentioned Fig. 1,
In order to avoid repeating, it will not be repeated here.
In addition, prior art using Page sort method or the Page sort method of subject-oriented carry out network node
Influence power sequence, using random surfer model, and make the ranking value converge to the threshold value of setting by iterative calculation to carry out
Influence power sequence.However, because the threshold value needs for making ranking value converge to setting using iterative calculation are repeatedly calculated, working as network
When interstitial content is excessive, computing cost is big.The method of the influence power of the determination network node of another embodiment of the present invention, Ke Yi
When network node number is excessive, the expense calculated is reduced.
Fig. 5 is the indicative flowchart of the method for the influence power of determination network node according to another embodiment of the present invention.
Fig. 5 method is performed by the device of the influence power of determination network node.Including：
510th, self-corresponding Page ranking value each to N number of network node carries out first time iteration, and first time iteration is iteration
Number of times limited number of time iteration set in advance；
520th, after according to first time iteration, N number of each self-corresponding Page ranking value of network node, by N number of network node point
For first set and second set, wherein, the corresponding Page ranking value of network node in second set is less than in first set
The corresponding Page ranking value of network node；
530th, self-corresponding Page ranking value each to N number of network node carries out second of iteration, in second of iteration, the
The iteration stopping threshold value of the corresponding Page ranking value of network node in one set is less than the network node correspondence in second set
Page ranking value iteration stopping threshold value；
540th, by after second of iteration, the size sequence of N number of each self-corresponding Page ranking value of network node is defined as N number of
The sequence of the influence power of network node.
It should be understood that network node can be website, cellphone subscriber, microblogging or wechat user, carrier gateway etc..Iteration time
Number limited number of time iteration set in advance, can be the iterations according to Six Degrees theory setting, such as 5 times.Can also be
Other iterationses, such as less than 10 times.In other words, first time iteration is that iterations limited number of time set in advance changes
Generation, in order to which the ranking value of the iteration result according to first time carries out the preliminary sequence of influence power, and as needed by the first collection
The network node of conjunction carries out more accurate sequence.Fig. 1 method can also include：In the first iteration, to N number of network section
The Page ranking value setting convergence outage threshold of point.After first time iteration, N number of each self-corresponding Page sequence of network node
Value, is divided into first set and second set by N number of network node, and Ke Yishi determines two set, example according to power-law distribution rule
In the Page ranking value after first time iteration, according to size order, preceding 20% network node is first set, rear 80%
Network node is second set.Different set can also be divided as needed.
Embodiments of the invention can set different convergence outage thresholds as desired for different network nodes, by
In for the small network node of convergence outage threshold the iterative calculation of less number of times can be carried out, therefore reduce opening for calculating
Pin.
The method that described above is the influence power of determination network node according to embodiments of the present invention, is retouched below in conjunction with the accompanying drawings
State the device of the influence power of determination network node according to embodiments of the present invention.
Fig. 6 is the schematic diagram of the device of the influence power of determination network node according to another embodiment of the present invention.
Fig. 6 device 600 includes：First iteration module 610, for carrying out the to each self-corresponding Page ranking value of N number of network node
An iteration, first time iteration is iterations limited number of time iteration set in advance；Determining module 620, for being changed according to first
Each self-corresponding Page ranking value of N number of network node obtained for module, is divided into first set and the second collection by N number of network node
Close, wherein, the corresponding Page ranking value of network node in second set is less than the corresponding pendant of network node in first set
Strange ranking value；Secondary iteration module 630, each self-corresponding Page row of N number of network node for being obtained to the first iteration module
Sequence value carries out second of iteration, in second of iteration, and the network node in the first set that determining module is determined is corresponding to wear
The iteration stopping threshold value of strange ranking value is less than the iteration stopping threshold value of the corresponding Page ranking value of network node in second set；
Determining module 620 is additionally operable to after second of iteration, and the size sequence of N number of each self-corresponding Page ranking value of network node is determined
For the sequence of the influence power of N number of network node.
Embodiments of the invention can set different convergence outage thresholds as desired for different network nodes, by
In for the small network node of convergence outage threshold the iterative calculation of less number of times can be carried out, therefore reduce opening for calculating
Pin.
The operation and function for determining the modules of the device 600 of the influence power of network node may be referred to above-mentioned Fig. 5's
Method, in order to avoid repeating, will not be repeated here.With reference to specific example, embodiments of the invention are described more fully.
Fig. 7 is the indicative flowchart of the method for the influence power of determination network node according to another embodiment of the present invention.
Fig. 7 method is the example of Fig. 5 method.In the embodiment of the network node corresponding diagram 5 in set Z in the present embodiment
The first set in the embodiment of the network node corresponding diagram 5 in set Z is not belonging in network node in two set, set Y
In network node；The iteration stopping threshold value of the corresponding Page ranking value of network node in first threshold correspondence first set,
The iteration stopping threshold value of the corresponding Page ranking value of network node in Second Threshold correspondence first set.First ranking value is used for
The initialization of ranking value vector, wherein the first ranking value of each network node is the element of ranking value vector；Second ranking value
The vector that N number of each self-corresponding Page ranking value of network node is constituted after vector correspondence first time iteration.Concrete implementation mode
It is as follows.
710th, iterations is set.
For example, by taking social networks as an example, iterations is set as into 5 according to Six Degrees theory, in other words, node is said
The transmission of influence power, which is no more than 5 times, can just reach all nodes.
720th, to the Page ranking value assignment of all-network node.
Specifically, the ranking value of identical first is set to all network nodes, for example, there will be network node to be entered as more
1。
730th, first threshold is set.
The first threshold is used for the fluctuation ratio with the influence power ranking value of network node compared with the fluctuation of influence power ranking value is determined
Justice is R=│ pk-pk+1│/pk。
740th, the influence power ranking value of subnetwork node is determined according to first threshold.
For example, using " model of levying a tax " formula 7.1:pk+1=β Mpk+(1-β)e/S (7.1)
Wherein, M is transition probability matrix, and the matrix is N N matrix；E is weight vectors, and the vector is the N number of network of correspondence
The N-dimensional vector of node；β is constant, for example, can be 0.15；pkAnd pk+1Respectively the ranking value of network node by kth time and
Ranking value vector after+1 iteration of kth, the vector is vectorial for the N-dimensional of the N number of network node of correspondence.S be N number of network node in
The number of the related network node of target topic.Element in weight vectors e depend on the node whether with the target topic phase
Close, be 1 if related, be otherwise 0.Specifically, using transition probability matrix M and weight vectors e to the first of network node
Ranking value vector is iterated calculating, and calculates after each iterative calculation the undulating value R of the ranking value of each network node.Will
Undulating value R is compared with first threshold：If the undulating value in 5 iterative calculation is not more than first threshold, stop to this
The calculating of network node.The ranking value of the network node is defined as influence power ranking value simultaneously, and the network node is determined
For the element in set X.If after 5 times iterate to calculate, the undulating value of the ranking value of network node is more than first threshold, then will
The network node is defined as element in set Y, and performs 750.
750th, the second ranking value vector of network node is determined.
Specifically, 5 iteration meters are carried out by multiple first ranking values vector to transition probability matrix and network node
Calculate, determine the second ranking value vector of network node, wherein the element in the second ranking value vector is the of each network node
Two ranking values.
760th, the set Z of low influence power network node in set Y is determined.
Specifically, tentatively sorted according to the size of the second ranking value of each network node in set Y, and according to row
Sequence result determines the set Z of low influence power network node.For example, according to power-law distribution that sequence is true in rear 80% network node
It is set to low influence power network node.
It is similar to 740 for the network node being not belonging in set Y in set Z, it is true according to first threshold and formula 7.1
The influence power ranking value of the fixed network node.770 are performed to the network node in set Z.
770th, Second Threshold is set for the network node in set Z.
Specifically, Second Threshold of the setting more than first threshold.Due to sort rear 80% network node be low influence
Power network node, therefore Second Threshold is not reduced the accuracy of final influence power sequence substantially more than first threshold.
780th, the influence power ranking value of network node in set Z is determined according to Second Threshold.
For example, according to formula 7.1, using transition probability matrix M and weight vectors e to the second ranking value of network node to
Amount is iterated calculating, and each undulating value of the ranking value of network node in set of computations Z after iterating to calculate every time.Should
Undulating value is compared with Second Threshold：If the undulating value is not more than Second Threshold, stop the calculating to the network node.Simultaneously
The ranking value of the network node is defined as to the influence power ranking value of network node in set Z.
790th, the influence power of all-network node is determined according to the influence power ranking value of all-network node.
Specifically, can be by network node according to influence because the influence power of the big network node of influence power ranking value is high
Power ranking value sorts to determine the influence power of all-network node from big to small.
Fig. 8 is the schematic diagram of the device of the influence power of determination network node according to another embodiment of the present invention.
Fig. 8 device 800 includes：Processor 810, memory 820, communication bus 830.Processor 810 is total by communication
Line 830 calls the code in memory 820, and is used for：Self-corresponding Page ranking value each to N number of network node is carried out for the first time
Iteration, first time iteration is iterations limited number of time iteration set in advance；After first time iteration, N number of network node is each
Self-corresponding Page ranking value, is divided into first set and second set by N number of network node, wherein, the network in second set
The corresponding Page ranking value of node is less than the corresponding Page ranking value of network node in first set；It is each to N number of network node
Self-corresponding Page ranking value carries out second of iteration, in second of iteration, and the network node in first set is corresponding to wear
The iteration stopping threshold value of strange ranking value is less than the iteration stopping threshold value of the corresponding Page ranking value of network node in second set；
After second of iteration, the size sequence of N number of each self-corresponding Page ranking value of network node is defined as the shadow of N number of network node
Ring the sequence of power.
The operation and function for determining the modules of the device 800 of the influence power of network node may be referred to above-mentioned Fig. 5's
Method, in order to avoid repeating, will not be repeated here.
In addition, the terms " system " and " network " are often used interchangeably herein.The terms " and/
Or ", only a kind of incidence relation for describing affiliated partner, represents there may be three kinds of relations, for example, A and/or B, can be with table
Show：Individualism A, while there is A and B, these three situations of individualism B.In addition, character "/" herein, before and after typicallying represent
Affiliated partner is a kind of relation of "or".
It should be understood that in embodiments of the present invention, " B " corresponding with A represents that B is associated with A, and B can be determined according to A.But
It should also be understood that determining that B is not meant to determine B only according to A according to A, B can also be determined according to A and/or other information.
Those of ordinary skill in the art are it is to be appreciated that the list of each example described with reference to the embodiments described herein
Member and algorithm steps, can be realized with electronic hardware, computer software or the combination of the two, in order to clearly demonstrate hardware
With the interchangeability of software, the composition and step of each example are generally described according to function in the above description.This
A little functions are performed with hardware or software mode actually, depending on the application-specific and design constraint of technical scheme.Specially
Industry technical staff can realize described function to each specific application using distinct methods, but this realization is not
It is considered as beyond the scope of this invention.
It is apparent to those skilled in the art that, for convenience of description and succinctly, foregoing description is
The specific work process of system, device and unit, may be referred to the corresponding process in preceding method embodiment, will not be repeated here.
, can be with several embodiments provided herein, it should be understood that disclosed systems, devices and methods
Realize by another way.For example, device embodiment described above is only schematical, for example, the unit
Divide, only a kind of division of logic function there can be other dividing mode when actually realizing, such as multiple units or component
Another system can be combined or be desirably integrated into, or some features can be ignored, or do not perform.In addition, shown or beg for
The coupling each other of opinion or direct-coupling or communication connection can be the INDIRECT COUPLINGs by some interfaces, device or unit
Or communication connection or electricity, mechanical or other forms are connected.
The unit illustrated as separating component can be or may not be it is physically separate, it is aobvious as unit
The part shown can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple
On NE.Some or all of unit therein can be selected to realize scheme of the embodiment of the present invention according to the actual needs
Purpose.
In addition, each functional unit in each embodiment of the invention can be integrated in a processing unit, can also
It is that unit is individually physically present or two or more units are integrated in a unit.It is above-mentioned integrated
Unit can both be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
Through the above description of the embodiments, it is apparent to those skilled in the art that the present invention can be with
Realized with hardware, or firmware is realized, or combinations thereof mode is realized.When implemented in software, can be by above-mentioned functions
It is stored in computer-readable medium or is transmitted as one or more instructions on computer-readable medium or code.Meter
Calculation machine computer-readable recording medium includes computer-readable storage medium and communication media, and wherein communication media includes being easy to from a place to another
Any medium of individual place transmission computer program.Storage medium can be any usable medium that computer can be accessed.With
Exemplified by this but it is not limited to：Computer-readable medium can include RAM, ROM, EEPROM, CD-ROM or other optical disc storages, disk
Storage medium or other magnetic storage apparatus or can be used in carrying or store with instruction or data structure form expectation
Program code and can by computer access any other medium.In addition.Any connection can be suitably turn into computer
Computer-readable recording medium.If for example, software is using coaxial cable, optical fiber cable, twisted-pair feeder, Digital Subscriber Line (DSL) or such as
The wireless technology of infrared ray, radio and microwave etc is transmitted from website, server or other remote sources, then coaxial electrical
The wireless technology of cable, optical fiber cable, twisted-pair feeder, DSL or such as infrared ray, wireless and microwave etc is included in affiliated medium
In fixing.As used in the present invention, disk (Disk) and dish (disc) are logical including compression laser disc (CD), laser disc, laser disc, numeral
With laser disc (DVD), floppy disk and Blu-ray Disc, the replicate data of the usual magnetic of which disk, and dish is then replicated with laser come optical
Data.Above combination above should also be as being included within the protection domain of computer-readable medium.
In a word, the preferred embodiment of technical solution of the present invention is the foregoing is only, is not intended to limit the present invention's
Protection domain.Within the spirit and principles of the invention, any modification, equivalent substitution and improvements made etc., should be included in
Within protection scope of the present invention.
Claims (8)
1. a kind of method for the influence power sequence for determining user, it is characterised in that including：
The first information and the second information are obtained, the first information is used for the mutual concern for indicating N number of user in social networks
Relation, message count of second information for indicating each user's issue in N number of user, and each user's issue
The theme of message；
According to the first information and second information, the corresponding N-dimensional weight vectors of target topic, the weight vectors are determined
In the i-th element be based on the first proportion and the second proportion determination, wherein, first proportion is：What the i-th user issued disappears
In breath, on the proportion shared by the message of the target topic, second proportion is：In the user of the i-th user concern,
The message package of issue contains on the proportion shared by the user of the message of the target topic, and i is 1 arbitrary integer into N；
The 3rd information is obtained, the 3rd information is used to indicate between the user with message forwarding relation in N number of user
Forward the quantity of message and the theme of the forwarding message；
According to the 3rd information, the corresponding transition probability matrix of the target topic, the of the transition probability matrix are determined
(j, k) element is to be based on：Kth user is forwarded by jth user, and on the quantity of the message of the target topic, the jth is used
The quantity of the corresponding forwarding association user in family, and the message on the target topic for forwarding association user to issue
What the magnitude relationship between quantity was determined, wherein, the forwarding association user is in N number of user, the message of issue is by institute
The forwarded over user of jth user is stated, j, k are 1 arbitrary integer into N；
According to the transition probability matrix and the weight vectors, determined using Page sort algorithm in the social networks,
Sequence of the N number of user in the influence power in the target topic field.
2. according to the method described in claim 1, it is characterised in that the i-th element in the weight vectors is by following public affairs
What formula was determined：
Ai=x (a/b)+(1-x) (c/d)
Wherein, AiIt is the i-th element in the weight vectors；X is real number set in advance, and 0≤x≤1；A sends out for the i-th user
The quantity of the message on the target topic of cloth；B is the message sum that i-th user issues；C is i-th user
In the user of concern, the message package of issue contains the quantity of the user of the message on the target topic；D is i-th user
The total number of users of concern.
3. method according to claim 1 or 2, it is characterised in that (j, k) element of the transition probability matrix is logical
Cross below equation determination：
Bjk=max (e, 1)/max (f, g)
Wherein, BjkFor (j, k) element of the transition probability matrix；E is what kth user was forwarded by jth user, on described
The quantity of the message of target topic；The quantity for the message on the target topic that f issues for the forwarding association user；g
For the quantity of the corresponding forwarding association user of the jth user.
4. method according to claim 1 or 2, it is characterised in that described according to the transition probability matrix and the power
Weight vector, the sequence of the influence power of N number of user is determined using Page sort algorithm, including：
According to the transition probability matrix and the weight vectors, N number of each self-corresponding Page ranking value of user is carried out
First time iteration, the first time iteration is iterations limited number of time iteration set in advance；
After the first time iteration, N number of user is divided into by N number of each self-corresponding Page ranking value of user
One set and second set, wherein, the corresponding Page ranking value of user in the second set is less than in the first set
The corresponding Page ranking value of user；
According to the transition probability matrix and the weight vectors, N number of each self-corresponding Page ranking value of user is carried out
Second of iteration, in second of iteration, the iteration stopping of the corresponding Page ranking value of user in the first set
Threshold value is less than the iteration stopping threshold value of the corresponding Page ranking value of user in the second set；
After second of iteration, the sequence of the size of N number of each self-corresponding Page ranking value of user is defined as described N number of
Sequence of the user in the influence power in the target topic field.
5. a kind of device for the influence power sequence for determining user, it is characterised in that including：
First acquisition module, for obtaining the first information and the second information, the first information is used to indicate the N in social networks
The mutual concern relation of individual user, second information is used to indicate the message count of each user's issue in N number of user, and
The theme of the message of each user's issue；
First determining module, for the first information obtained according to first acquisition module and second information, really
The i-th element in the corresponding N-dimensional weight vectors of the theme that sets the goal, the weight vectors is true based on the first proportion and the second proportion
Fixed, wherein, first proportion is：In the message of i-th user issue, on the ratio shared by the message of the target topic
Heavy, second proportion is：In the user of the i-th user concern, the message package of issue contains disappearing on the target topic
Proportion shared by the user of breath, i is 1 arbitrary integer into N；
Second acquisition module, for obtaining the 3rd information, the 3rd information is used to indicate in N number of user to turn with message
The quantity of message and the theme of the forwarding message are forwarded between the user of hair relation；
Second determining module, for the 3rd information obtained according to second acquisition module, determines the target topic
Corresponding transition probability matrix, (j, k) element of the transition probability matrix is to be based on：Kth user is forwarded by jth user
, on the quantity of the message of the target topic, the quantity of the corresponding forwarding association user of the jth user, and it is described
What the magnitude relationship between the quantity of the message on the target topic of forwarding association user issue was determined, wherein, it is described
Association user is forwarded in N number of user, the message of issue is by the forwarded over user of the jth user, and j, k are 1 into N
Arbitrary integer；
3rd determining module, for according to the transition probability matrix and the weight vectors, being determined using Page sort algorithm
In the social networks, sequence of the N number of user in the influence power in the target topic field.
6. device according to claim 5, it is characterised in that the i-th element in the weight vectors is by following public affairs
What formula was determined：
Ai=x (a/b)+(1-x) (c/d)
Wherein, AiIt is the i-th element in the weight vectors；X is real number set in advance, and 0≤x≤1；A sends out for the i-th user
The quantity of the message on the target topic of cloth；B is the message sum that i-th user issues；C is i-th user
In the user of concern, the message package of issue contains the quantity of the user of the message on the target topic；D is i-th user
The total number of users of concern.
7. the device according to claim 5 or 6, it is characterised in that (j, k) element of the transition probability matrix is logical
Cross below equation determination：
Bjk=max (e, 1)/max (f, g)
Wherein, BjkFor (j, k) element of the transition probability matrix；E is what kth user was forwarded by jth user, on described
The quantity of the message of target topic；The quantity for the message on the target topic that f issues for the forwarding association user；g
For the quantity of the corresponding forwarding association user of the jth user.
8. the device according to claim 5 or 6, it is characterised in that the 3rd determining unit is specifically for according to described
Transition probability matrix and the weight vectors, first time iteration, institute are carried out to N number of each self-corresponding Page ranking value of user
First time iteration is stated for iterations limited number of time iteration set in advance；After the first time iteration, N number of user is each
Self-corresponding Page ranking value, is divided into first set and second set by N number of user, wherein, in the second set
The corresponding Page ranking value of user is less than the iteration stopping threshold value of the corresponding Page ranking value of user in the first set；Root
According to the transition probability matrix and the weight vectors, N number of each self-corresponding Page ranking value of user is carried out second
Iteration, in second of iteration, the iteration stopping threshold value of the corresponding Page ranking value of user in the first set is small
The iteration stopping threshold value of the corresponding Page ranking value of user in the second set；After second of iteration, the N
The size sequence of individual each self-corresponding Page ranking value of user is defined as influence of the N number of user in the target topic field
The sequence of power.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410294986.8A CN104092567B (en) | 2014-06-26 | 2014-06-26 | Determine the method and apparatus of the influence power sequence of user |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410294986.8A CN104092567B (en) | 2014-06-26 | 2014-06-26 | Determine the method and apparatus of the influence power sequence of user |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104092567A CN104092567A (en) | 2014-10-08 |
| CN104092567B true CN104092567B (en) | 2017-10-27 |
Family
ID=51640249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410294986.8A Active CN104092567B (en) | 2014-06-26 | 2014-06-26 | Determine the method and apparatus of the influence power sequence of user |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104092567B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105306540A (en) * | 2015-09-24 | 2016-02-03 | 华东师范大学 | Method for obtaining top k nodes with maximum influence in social network |
| CN108304395B (en) * | 2016-02-05 | 2022-09-06 | 北京迅奥科技有限公司 | Webpage cheating detection |
| CN106874428B (en) * | 2017-01-23 | 2021-05-07 | 北京航空航天大学 | Method and device for selecting key node of information propagation |
| CN108470344A (en) * | 2017-02-23 | 2018-08-31 | 南宁市富久信息技术有限公司 | The Sobel edge detection methods of adaptive threshold |
| CN107240029B (en) * | 2017-05-11 | 2023-03-31 | 腾讯科技（深圳）有限公司 | Data processing method and device |
| CN108183956B (en) * | 2017-12-29 | 2020-05-12 | 武汉大学 | Method for extracting key path of propagation network |
| CN109272225A (en) * | 2018-09-07 | 2019-01-25 | 大连海事大学 | A kind of Bug tracing system tester importance ranking method |
| CN109190058A (en) * | 2018-10-15 | 2019-01-11 | 北京字节跳动网络技术有限公司 | Method and apparatus for handling information |
| CN109800289B (en) * | 2019-02-26 | 2022-05-17 | 合肥工业大学 | Method and system for identifying network user, and method and system for shielding network information |
| CN111988129B (en) * | 2019-05-21 | 2022-07-01 | 中移(苏州)软件技术有限公司 | Influence maximization data set processing method, device and system |
| CN112395510A (en) * | 2019-08-15 | 2021-02-23 | 北京京东尚科信息技术有限公司 | Method and device for determining target user based on activity |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6285999B1 (en) * | 1997-01-10 | 2001-09-04 | The Board Of Trustees Of The Leland Stanford Junior University | Method for node ranking in a linked database |
| CN101976245A (en) * | 2010-10-09 | 2011-02-16 | 吕琳媛 | Sequencing method of node importance in network |
| CN102945279A (en) * | 2012-11-14 | 2013-02-27 | 清华大学 | Evaluating method and device of influence effect of microblog users |
| CN103309957A (en) * | 2013-05-28 | 2013-09-18 | 华东师范大学 | Social network expert locating method introducing levy flight |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080103907A1 (en) * | 2006-10-25 | 2008-05-01 | Pudding Ltd. | Apparatus and computer code for providing social-network dependent information retrieval services |
-
2014
- 2014-06-26 CN CN201410294986.8A patent/CN104092567B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6285999B1 (en) * | 1997-01-10 | 2001-09-04 | The Board Of Trustees Of The Leland Stanford Junior University | Method for node ranking in a linked database |
| CN101976245A (en) * | 2010-10-09 | 2011-02-16 | 吕琳媛 | Sequencing method of node importance in network |
| CN102945279A (en) * | 2012-11-14 | 2013-02-27 | 清华大学 | Evaluating method and device of influence effect of microblog users |
| CN103309957A (en) * | 2013-05-28 | 2013-09-18 | 华东师范大学 | Social network expert locating method introducing levy flight |
Non-Patent Citations (3)
| Title |
|---|
| "Analysis of user influence in social network based on behavior and relationship";Huang Yulan，Li Ling;《IEEE》;20130816;682-686 * |
| "The pagerank citation ranking：Bring Order to the Web";Page L;《Stanford Digital Libraries》;19991231;全文 * |
| "基于PageRank的社交网络用户实时影响力研究";陈少钦;《中国优秀硕士学位论文全文数据库信息科技辑》;20130731;I139-205 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104092567A (en) | 2014-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104092567B (en) | Determine the method and apparatus of the influence power sequence of user | |
| CN104331411B (en) | The method and apparatus of recommended project | |
| CN104657418B (en) | A kind of complex network propagated based on degree of membership obscures corporations' method for digging | |
| Zhang et al. | The exact solution of the mean geodesic distance for Vicsek fractals | |
| Hu et al. | Evaluating node importance with multi-criteria | |
| CN112052404A (en) | Group discovery method, system, device and medium for multi-source heterogeneous relation network | |
| CN107222410B (en) | Method, device, terminal and computer readable storage medium for link prediction | |
| Yeh | A sequential decomposition method for estimating flow in a multi-commodity, multistate network | |
| CN108259195A (en) | The determining method and system of the coverage of anomalous event | |
| Xiao et al. | Improving Rete algorithm to enhance performance of rule engine systems | |
| CN117560275B (en) | Root cause positioning method and device for micro-service system based on graphic neural network model | |
| CN114862449A (en) | Method and device for calculating unique natural person identifier, electronic equipment and storage medium | |
| CN113312560B (en) | Group detection method and device and electronic equipment | |
| CN110413722A (en) | Address choice method, apparatus and non-transient storage medium | |
| CN113868537A (en) | Recommendation method based on multi-behavior session graph fusion | |
| CN109063024A (en) | A kind of social platform user force calculation method and device | |
| CN109033603B (en) | Intelligent substation secondary system simulation method based on source flow path chain | |
| CN109255433B (en) | Community detection method based on similarity | |
| Qiao et al. | Product module network modeling and evolution analysis | |
| CN115563310A (en) | Method, device, equipment and medium for determining key service node | |
| CN113010747B (en) | Information matching method, device, equipment and storage medium | |
| CN110223125B (en) | User position obtaining method under node position kernel-edge profit algorithm | |
| CN107392784A (en) | The adjustable overlapping Combo discovering method propagated based on label | |
| CN116739470B (en) | Network analysis method and device | |
| CN103957012A (en) | Method and device for compressing DFA matrix |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201103 Address after: 625, room 269, Connaught platinum Plaza, No. 518101, Qianjin Road, Xin'an street, Shenzhen, Guangdong, Baoan District Patentee after: SHENZHEN SHANGGE INTELLECTUAL PROPERTY SERVICE Co.,Ltd. Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee before: HUAWEI TECHNOLOGIES Co.,Ltd. |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20201126 Address after: 213000 No. 11 Qingyang North Road, Tianning District, Changzhou, Jiangsu Patentee after: Changzhou Hong quantity Electronic Technology Co.,Ltd. Address before: 625, room 269, Connaught platinum Plaza, No. 518101, Qianjin Road, Xin'an street, Shenzhen, Guangdong, Baoan District Patentee before: SHENZHEN SHANGGE INTELLECTUAL PROPERTY SERVICE Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220822 Address after: Tianning District Hehai road 213000 Jiangsu city of Changzhou province No. 9 Patentee after: Changzhou Tianning Communication Technology Industrial Park Co.,Ltd. Address before: 213000 No. 11 Qingyang North Road, Tianning District, Changzhou City, Jiangsu Province Patentee before: Changzhou Hong quantity Electronic Technology Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221221 Address after: Tianning District Hehai road 213000 Jiangsu city of Changzhou province No. 9 Patentee after: Changzhou Hengtang Technology Industry Co.,Ltd. Address before: Tianning District Hehai road 213000 Jiangsu city of Changzhou province No. 9 Patentee before: Changzhou Tianning Communication Technology Industrial Park Co.,Ltd. |
|
| TR01 | Transfer of patent right |