CN109800351A - High-impact usage mining method in microblogging specific topics - Google Patents
High-impact usage mining method in microblogging specific topics Download PDFInfo
- Publication number
- CN109800351A CN109800351A CN201811629337.3A CN201811629337A CN109800351A CN 109800351 A CN109800351 A CN 109800351A CN 201811629337 A CN201811629337 A CN 201811629337A CN 109800351 A CN109800351 A CN 109800351A
- Authority
- CN
- China
- Prior art keywords
- user
- topic
- forwarding
- calculates
- blog article
- 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
Abstract
The present invention discloses a kind of high-impact usage mining method in microblogging specific topics, includes the following steps: that (10) crawl microblog topic: using crawlers, crawling the related truthful data of specific microblog topic；(20) network divides: splitting the network into user and forwards relational network and user's blog article forwarding relational network, different time forwarding is to the contribution degree for being forwarded user force between analyzing user；(30) user's topic forwarding influence power calculates: forwarding relational network based on user, calculates user's topic and forward influence power；(40) user's sole mass calculates: forwarding relational network based on user's blog article, calculates user's sole mass；(50) user's topic information transmission capacity calculates: calculating user's topic information transmission capacity；(60) user's topic propagation effect power, and descending arrangement output high-impact usage mining: is calculated.High-impact usage mining method of the invention, precision is high, timeliness is good.
Description
Technical field
The invention belongs to social network data processing technology field, the microblogging that especially a kind of precision is high, timeliness is good is special
Determine high-impact usage mining method in topic.
Background technique
The rise of social networks changes the mode of people's acquisition of information and friend-making, brings huge change to people's life
Become.Platform is propagated as the expression of the individual freedom of representative with the novel public topic exchanged using microblogging, increasingly to social influence day
Deep, one side microblogging makes media event fast propagation and causes public extensive concern, another aspect government, school and all kinds of
Mechanism is also actively engaged in microblogging, so that the actual influence power ever more powerful of microblogging.The massive information that microblogging is contained also makes it
Become the hot spot of research.
Microblog users influence power analyzes one of the important content studied as social networks, monitors to network marketing, public sentiment
Equal numerous areas all produce important guidance and development function.Microblog users influence power can pass through the social network between user
Network activity embodies, and behavior and the viewpoint etc. for showing as user are influenced the phenomenon that changing by other people.Different user performance
Influence power out contains huge commercial value and social value, analyzes user force based on microblog and excavates height
Influence power user understands social network it can be found that the mechanism of transmission and rule of topic information, monitor the public feelings information of social networks
The social property of network recommends the high quality good friend of common interest, and the government affairs of business development and government bodies to enterprise-like corporation are determined
Plan etc. is all of great significance.
At present, researchers are mainly based upon two class of network structure and user behavior to the analysis method of user force.
Analysis method based on network structure is the topological relation structure formed during social activity according to people, with network topological diagram
Knowledge measures the importance of node, but the topological relation of the building of the method based on network structure is more single, underuse use
Interactive information between family.Analysis method based on user behavior is to utilize the behaviors such as concern, publication, forwarding, the comment of user
Data assess the influence power of behavior promoter and message sender, but due to secret protection etc., disclosed user behavior
Data are not very complete, and can not only embody the true horizon of user force completely with user behavior.
In short, problem of the existing technology is: it is not high enough to the computational accuracy of microblogging specific topics user force, when
Effect property is not strong enough, and there is " waterborne troops " phenomenon of the brush list within the short time, cause network public-opinion monitoring and marketing inaccuracy,
Not in time.
Summary of the invention
The purpose of the present invention is to provide a kind of high-impact usage mining method in microblogging specific topics, precision is high, when
Effect property is good.
The technical solution for realizing the aim of the invention is as follows:
A kind of high-impact usage mining method in microblogging specific topics, includes the following steps:
(10) it crawls microblog topic: using crawlers, crawling the related truthful data of specific microblog topic；
(20) network divides: according to the interactive relationship between user and topic, splitting the network into user and forwards network of personal connections
Network and user's blog article forward relational network, and based on different time forwarding between partitioning network analysis user to being forwarded user's shadow
Ring the contribution degree of power；
(30) user's topic forwarding influence power calculates: relational network is forwarded based on user, according to user activity and user
Time of the act calculates user's topic and forwards influence power；
(40) user's sole mass calculates: relational network is forwarded based on user's blog article, according to user's blog article quality and user
Time of the act calculates user's sole mass；
(50) user's topic information transmission capacity calculates: forwarding influence power and user's sole mass, meter according to user's topic
Calculate user's topic information transmission capacity；
(60) high-impact usage mining: user's topic information transmission capacity and user and the microblog topic degree of association are carried out
User's topic propagation effect power, and descending arrangement output is calculated in linear fusion.
Compared with prior art, the present invention its remarkable advantage are as follows:
1, precision is high: be compared to single network structure and user behavior data, the present invention is based on topic propagation law and
It analyzes to mechanism, more fine granularity user and is participating in forwarding influence power and user's sole mass in topic communication process, body
Reveal the relay range of the topic information propagation capabilities of user and influences intensity, meanwhile, in conjunction with the degree of association of user and topic, more
Add the influence power for accurately analyzing topic associated user.
2, timeliness is good: having quantified the forwarding of user's different time to being forwarded the influence power contribution degree of user, and has been fused to
User force is calculated in two kinds of forwarding relational networks, is conducive to new high-impact user and floats, precipitate outmoded pseudo- Gao Ying
Ring power user.
3, topic relevance is good: being based on correlation analysis user force between user and topic, realizes and excavate very
The high-impact user of face topic continued interest can effectively limit " waterborne troops " non-body by brush list in the short time merely
Reveal the behavior to the concern of topic continuous interest.
The present invention is described in further detail with reference to the accompanying drawings and detailed description.
Detailed description of the invention
Fig. 1 is the flow chart of high-impact usage mining method in microblogging specific topics of the present invention.
Fig. 2 is the flow chart that user's topic forwarding influence power calculates in Fig. 1.
Fig. 3 is the flow chart of high-impact usage mining method in Fig. 1.
Fig. 4 is capture rate experimental result picture provided in an embodiment of the present invention.
Fig. 5 is accuracy rate experimental result picture provided in an embodiment of the present invention.
Fig. 6 is recall rate experimental result picture provided in an embodiment of the present invention.
Specific embodiment
As shown in Figure 1, high-impact usage mining method in microblogging specific topics of the present invention, includes the following steps:
(10) it crawls microblog topic: using crawlers, crawling the related truthful data of specific microblog topic；
As embodiment, the crawlers of microblog topic data can be voluntarily write, are crawled from microblog true micro-
Related data of rich topic, including content of microblog, forwarding content, comment content etc..Existing crawlers can also be used and realize spy
Determine microblog topic related data to crawl.Because being the prior art, do not repeat herein.
(20) network divides: according to the interactive relationship between user and topic, splitting the network into user and forwards network of personal connections
Network and user's blog article forward relational network, and based on different time forwarding between partitioning network analysis user to being forwarded user's shadow
Ring the contribution degree of power；
During (20) network divides, the contribution degree of quantization user's different time forwarding is specifically, be calculated as follows
Different time forwarding is to the contribution degree for being forwarded user force between user:
Wherein, O1(u) indicate that user u issues topic blog article set,For user v forward i-th blog article of user u when
Between poor (time difference is smaller, bigger to the influence power contribution degree for being forwarded user u), λ be control rate of decay parameter, be arranged λ
=11 (h, hours).
ri(V, u) represent i-th blog article whether user v forwards user u.E is natural constant (e > 0)；
User u may issue a plurality of blog article in relation to topic in topic, and another user v may forward user u a plurality of rich
Text, therefore use riRepresent the forwarding relationship of user v Yu i-th blog article of user u.
It is expressed as follows:
Use is constructed respectively by the relationship between blog article in the forwarding relationship between topic participating user and user and topic
Family forwards relational network and user's blog article to forward relational network.
Use riRepresent the forwarding relationship of user v Yu i-th blog article of user u because user u may be issued in topic it is a plurality of
Blog article in relation to topic, another user v may forward a plurality of blog article of user u, wr(v, u) indicates that user v propagates user's u information
The forwarding contribution degree of ability, i.e. user v forward time of the act to embody forwarding time pair the forwarding weight of user u, user
It is forwarded the contribution degree of user force, i.e., the forwarding of time is more advantageous to the propagation for being forwarded information earlier, to being forwarded use
The contribution degree of family influence power is higher.
User forwards relational network to be expressed as G1=(V1,E1,W1), user's blog article forwarding relational network is expressed as G2=(V2,
E,W2).Wherein, V1To participate in topic whole user, E1The oriented line set of relationship, W are forwarded between user1For turning for directed edge
Weight is sent out, which is portrayed by forwarding time of the act and user activity, V2It is user and blog article two types node collection
It closes, i.e. V2=(VB∪VU), VB={ b1,b2,…,bmRepresent blog article set, VU={ u1,u2,…,unRepresent user node collection
It closes, E=EU→B∪EB→UIt is two class line sets,It is user to one group of side (user of blog article node
Forward blog article),It is blog article to one group of side of user node (blog article owning user), W2For user
The weight for forwarding side in relational network, be divided into two classes: the side right weight and blog article node of user node to blog article node are saved to user
The side right weight of point.
(30) forwarding of user's topic influences force calculating: relational network is forwarded based on user, according to user activity and use
Family time of the act calculates user's topic and forwards influence power；
As shown in Fig. 2, (30) user's topic forwarding influence force calculating step includes:
(31) user activity calculates: user activity is calculated as follows,
Wherein, npostIt (u) is that user issues blog article quantity, n in period TrepostIt (u) is user in period T
Quantity is forwarded, T is time segment length；
The liveness of user u is expressed as a (u), by the dispatch frequency of user u whithin a period of time and forwarding frequency come amount
Change the liveness that user u promotes information to propagate in topic.
(32) user's topic forwarding influence power calculates: user's topic forwarding influence power is calculated as follows,
Wherein, R (u) is the forwarding influence power of user u, O2(u) gather for the forwarding user of user u, R (v) is user v's
Influence power is forwarded, out (v) is the forwarding that user v is directed toward other users, and c is damped coefficient, is usually arranged as empirical value 0.85.
The influence of user's information propagation capabilities in topic is measured by calculating the forwarding influence power of topic participating user
Crowd's range, comprehensive PageRank thought and introducing forwarding time property forward the influence factor of relationship strength to user.
(40) user's sole mass calculates: relational network is forwarded based on user's blog article, according to user's blog article quality and user
Time of the act calculates user's sole mass；
(40) user's sole mass calculates step specifically, user's sole mass is calculated as follows:
Wherein, O4(b) be user u all blog article set,
wb(u) contribution degree that user's u sole mass is calculated for the blog article b of user u,
wb(u)=Nb/Nc,
NbFor total forwarding number of the blog article b of user u, NcFor the total forwarding number of all blog articles of user u.
User forwards behavior by the value driving of user itself, and user quality is higher in topic, the influence to other users
Degree is bigger, embodies the influence intensity of the information propagation capabilities of user.In microblog topic communication process, because of user itself
Quality is mainly embodied by the quality that user issues blog article.
Firstly, user u issues blog article b (b ∈ V in user's blog article forwarding relational networkB) quality be denoted as Blogb(u),
The forwarding relationship side right of user to blog article node, which is reseted, is set to wr(v, u) embodies forwarding time difference and propagates range to blog article and win
The influence of text forwarding relationship strength, uses for reference PageRank thought and calculates Blogb(u), calculation method is as follows:
Wherein, O3(b) be blog article b forwarding user set, Q (v) is the sole mass of user v, and out (v) is that user v refers to
Forwarding to other users.
(50) user's topic information transmission capacity calculates: forwarding influence power and user's sole mass, meter according to user's topic
Calculate user's topic information transmission capacity；
It is described that ((50) user topic information transmission capacity calculates step specifically, user's topic information biography is calculated as follows
Broadcast ability:
Spread (u)=α1×R(u)+α2×Q(u)
Wherein, α1It is that user forwards influence power proportion, α2It is user's sole mass proportion, α is set1=α2=
0.5, indicate that the forwarding influence power of user and user's sole mass are of equal importance to user information transmission capacity.
R (u) is forwarding influence power of the user u in topic participation process, and Q (u) is user u table in topic participation process
The sole mass revealed.
Two angles of influence power and user's sole mass are forwarded to calculate user information by the method for linear fusion from user
Transmission capacity Spread (u), the influence for embodying microblog users influence power respectively propagate range and influence user's intensity.
(60) high-impact usage mining: user's topic information transmission capacity and user and the microblog topic degree of association are carried out
User's topic propagation effect power, and descending arrangement output is calculated in linear fusion.
As shown in figure 3, (60) the high-impact usage mining step includes:
(61) user and microblog topic calculation of relationship degree: being calculated as follows user and the microblog topic degree of association,
In formula, the theme probability distribution V of topic collection of documenttopic, vector VtopicKL distance D between VuKL(Vu||
Vtopic),
The theme probability distribution V of customer documentation setu,
The theme probability distribution V of microblog topic collection of documenttopic,
In formula,WithBe respectively user u collection of document and microblog topic collection of document generate theme i probability,
And
Since KL divergence does not have symmetry, the symmetry of the degree of association is not consistent between user and background topic.In order to
So that formula meets symmetry, convenient for description user and the background topic degree of association and KL value corresponding relationship, formula has been carried out turn
It changes, defines degree of association S (u, topic) between user u and background topic.
(62) user's topic propagation effect power calculates: user's topic propagation effect power is calculated as follows,
TSRank (u)=Spread (u) × S (u, topic)；
(63) it high-impact usage mining: is arranged according to user's topic propagation effect power descending, obtains influence power ranking and lean on
Preceding high-impact user.
The case history blog article set in the period is corresponded to by the related all blog article set of analysis microblog topic and user, is counted
Calculate the degree of association between user and microblog topic.Microblogging short text is not suitable for for LDA topic model, by the way of polymerization
By in microblog topic each blog article and other users comment on its content and the reply commented on these of original publisher
Multiple single blog article set are aggregated into expand microblogging short text, then, LDA topic model identification microblog topic is literary using improving
The theme probability distribution V of shelves settopicWith the theme probability distribution V of customer documentation setu, pass through relative entropy (Kullback-
Leibler Divergence, KLD) calculate VtopicAnd VuThe distance between vector, value is bigger, illustrates more dissimilar.
High-impact usage mining method is in topic provided by the invention, and the information based on topic participating user propagates energy
Power and user and the user force calculation method for participating in the degree of association between topic.Firstly, being crawled very by web crawlers technology
Real microblog topic data (the relevant all users of topic and the relevant all blog articles of topic and forwarding, comment etc. between them
Kinds of relationships data), then, based on microblog topic, analysis user forwards time of the act, further constructs user's forwarding and uses
Family blog article forwards two kinds of topics to forward relational network, predicts user's topic information transmission capacity, analyzes individual subscriber historical weibo
With background topic microblogging content of text, the relevance between user and background topic is excavated, finally, comprehensively considering user's topic letter
The influence power that relevance between transmission capacity and user and background topic calculates microblog topic participating user is ceased, is found out in topic
The high-impact user of top-k.The present invention is a kind of user force analysis method towards microblog topic, synthetic user behavior
Timeliness, user activity and user forward relational network to calculate user and forward influence power, embody influence power in topic
Range, synthetic user time of the act and user's blog article forwarding network are propagated to calculate user's sole mass, embodies influence power
The influence intensity to user is propagated in topic to find out by the degree of association between analysis topic participating user and background topic
To the high-impact user that topic is really interested and gives more sustained attention, existing research method is compared, the present invention not only increases
The computational accuracy of user force and the timeliness for embodying influence power can more be found out in the topic time and really feel emerging to topic
The long lasting effect power user of interest can exclude the pseudo- high-impact user by brush list in " waterborne troops " short time to a certain extent,
To provide technical support for the problems such as monitoring of solution network public-opinion and marketing.
Effectiveness of the invention is verified in a manner of specific embodiment below.
1, data set
The data set that the present invention is embodied is crawler to three different types of true microblog topic data: " 2017NBA
Finals ", " Xu Yuyu case " and " breaking off a friendship door along rich green hand's data ", each topic data include that topic blog article content, forwarding are rich
Text, comment blog article, friend relation etc., the details of data set see the table below 1:
Table 1
In addition, Experimental Hardware environment: Ubuntu16.04, i7-3520M processor, memory 8G, hard disk 300GB, code fortune
Row environment: Hadoop2.6.0, python2.7, MATLAB2016a, Pycharm.
2, experimental result
We have chosen three capture rate (Catch Ratio, CR), accuracy rate and recall rate evaluation metrics to assess experiment
As a result, CR is the ratio of the information of physical presence in the information detected by Top-k user and network, accuracy rate
(precision) authenticity for excavating influence power user in topic is embodied, recall rate R (Recall) is embodied in microblog topic and influenced
The abundant degree of the excavation of power user, accuracy rate and recall rate we by the way of cross validation, the crossing number of selection is 3.Choosing
Three classical user force parser WBRank, TwitterRank and PageRank algorithms and inventive algorithm are taken to exist
It is run on three true topic datas, wherein WBRank algorithm considers that user behavior is analyzed, and TwitterRank is calculated
Method is the classic algorithm based on topic and network structure, and PageRank algorithm is classical sort algorithm.
The CR that experiment gives the lower four kinds of algorithms of Top-10, Top-20, Top-30, Top-50, Top-80 and Top-100 refers to
Experimental result is marked, sees Fig. 4.It is found through comparing, with increasing for powerful number of users is chosen, the method for the present invention CR index is equal
Better than other three kinds of methods, illustrate the better effect of powerful user in inventive algorithm identification microblog topic.
Experiment equally provide Average Accuracy under three topics and recall rate Top10, Top20, Top30, Top50,
Experimental result under Top-80 and Top-100 is shown in attached drawing 5 and attached drawing 6.It is found through comparing, likewise as the increasing for choosing number of users
More, the accuracy rate and recall rate of the method for the present invention are substantially better than other three kinds of methods, show that the method for the present invention more can fully be dug
The high-impact user under microblogging specific topics is excavated, moreover, other three kinds of methods that compare, the method for the present invention also have preferably
Efficiency of algorithm.Wherein, because what PageRank algorithm mainly took is more single network structure sort method, with microblogging spy
Property combine it is less, so experiment effect and the experiment effect gap of other three kinds of methods are more apparent.
3, experiment conclusion
Pass through the experiment on three true microblog topic data sets, it is found that be compared to above-mentioned three kinds classical use
Family influence power analysis method, it was demonstrated that the method for the present invention forwards two kinds of forwarding relational networks in building user forwarding and user's blog article
In structure basis, time of the act is forwarded to analyze the feasibility of user information transmission capacity in conjunction with user, moreover, introducing topic
The degree of association between participating user and background topic can excavate more accurately real high-impact user in topic, the present invention
Method all shows better effect in terms of accuracy rate, recall rate and efficiency of algorithm, meanwhile, inventive algorithm is in certain journey
It can contain the bad user behavior by " waterborne troops " brush list on degree.
Claims (6)
1. a kind of high-impact usage mining method in microblogging specific topics, which comprises the steps of:
(10) it crawls microblog topic: using crawlers, crawling the related truthful data of specific microblog topic；
(20) network divides: according to the interactive relationship between user and topic, split the network into user forward relational network and
User's blog article forwards relational network, and based on different time forwarding between partitioning network analysis user to being forwarded user force
Contribution degree；
(30) user's topic forwarding influence power calculates: relational network is forwarded based on user, according to user activity and user behavior
Time calculates user's topic and forwards influence power；
(40) user's sole mass calculates: relational network is forwarded based on user's blog article, according to user's blog article quality and user behavior
Time calculates user's sole mass；
(50) user's topic information transmission capacity calculates: forwarding influence power and user's sole mass according to user's topic, calculates and use
Family topic information transmission capacity；
(60) high-impact usage mining: user's topic information transmission capacity and user and the microblog topic degree of association are carried out linear
User's topic propagation effect power, and descending arrangement output is calculated in fusion.
2. method for digging according to claim 1, which is characterized in that during (20) network divides, quantization user is different
The contribution degree of time forwarding is specifically, be calculated as follows tribute of the different time forwarding to user force is forwarded between user
Degree of offering:
Wherein, O1(u) indicate that user u issues topic blog article set,The time difference of i-th blog article of user u is forwarded for user v,
λ is the parameter for controlling rate of decay, ri(v, u) represents i-th blog article whether user v forwards user u.
3. method for digging according to claim 2, which is characterized in that (30) user's topic forwarding influences force and calculates
Step includes:
(31) user activity calculates: user activity is calculated as follows,
Wherein, npostIt (u) is that user issues blog article quantity, n in period TrepostIt (u) is forwarding of the user in period T
Quantity, T are time segment length；
(32) user's topic forwarding influence power calculates: user's topic forwarding influence power is calculated as follows,
Wherein, R (u) is the forwarding influence power of user u, O2(u) gather for the forwarding user of user u, R (v) is the forwarding of user v
Influence power, out (v) are the forwarding that user v is directed toward other users, and c is damped coefficient.
4. method for digging according to claim 3, which is characterized in that it is specific that (40) user's sole mass calculates step
For user's sole mass is calculated as follows:
Wherein, O4(b) be user u all blog article set,
wb(u) contribution degree that user's u sole mass is calculated for the blog article b of user u,
wb(u)=Nb/Nc,
NbFor total forwarding number of the blog article b of user u, NcFor the total forwarding number of all blog articles of user u.
5. method for digging according to claim 4, which is characterized in that ((50) user topic information transmission capacity meter
Step is calculated specifically, user's topic information transmission capacity is calculated as follows:
Spread (u)=α1×R(u)+α2×Q(u)
Wherein, α1It is that user forwards influence power proportion, α2It is user's sole mass proportion, R (u) is user u in topic
Forwarding influence power in participation process, Q (u) are the sole mass that user u is shown in topic participation process.
6. method for digging according to claim 5, which is characterized in that (60) the high-impact usage mining step packet
It includes:
(61) user and microblog topic calculation of relationship degree: being calculated as follows user and the microblog topic degree of association,
In formula, the theme probability distribution V of topic collection of documenttopic, vector VtopicKL distance D between VuKL(Vu||
Vtopic),
The theme probability distribution V of customer documentation setu,
The theme probability distribution V of microblog topic collection of documenttopic,
In formula, pu iAnd ptopic iBe respectively user u collection of document and microblog topic collection of document generate theme i probability, and
(62) user's topic propagation effect power calculates: user's topic propagation effect power is calculated as follows,
TSRank (u)=Spread (u) × S (u, topic)；
(63) it high-impact usage mining: is arranged according to user's topic propagation effect power descending, it is in the top to obtain influence power
High-impact user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811629337.3A CN109800351A (en) | 2018-12-29 | 2018-12-29 | High-impact usage mining method in microblogging specific topics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811629337.3A CN109800351A (en) | 2018-12-29 | 2018-12-29 | High-impact usage mining method in microblogging specific topics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109800351A true CN109800351A (en) | 2019-05-24 |
Family
ID=66558009
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811629337.3A Pending CN109800351A (en) | 2018-12-29 | 2018-12-29 | High-impact usage mining method in microblogging specific topics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109800351A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110209962A (en) * | 2019-06-12 | 2019-09-06 | 合肥工业大学 | The acquisition methods and system of theme level high-impact user |
| CN110910176A (en) * | 2019-11-27 | 2020-03-24 | 上海风秩科技有限公司 | Critical consumer recruitment method, apparatus, computer device and readable storage medium |
| CN110929172A (en) * | 2019-11-27 | 2020-03-27 | 中科曙光国际信息产业有限公司 | Information selection method and device, electronic equipment and readable storage medium |
| CN111241420A (en) * | 2020-01-10 | 2020-06-05 | 云境商务智能研究院南京有限公司 | Recommendation method based on social network information diffusion perception |
| CN113781250A (en) * | 2020-09-14 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Social media information propagation evaluation method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116605A (en) * | 2013-01-17 | 2013-05-22 | 上海交通大学 | Method and system of microblog hot events real-time detection based on detection subnet |
| CN103179198A (en) * | 2012-11-02 | 2013-06-26 | 中国人民解放军国防科学技术大学 | Topic influence individual digging method based on relational network |
| CN103678474A (en) * | 2013-09-24 | 2014-03-26 | 浙江大学 | Method for acquiring large number of hot topics fast in social network |
-
2018
- 2018-12-29 CN CN201811629337.3A patent/CN109800351A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103179198A (en) * | 2012-11-02 | 2013-06-26 | 中国人民解放军国防科学技术大学 | Topic influence individual digging method based on relational network |
| CN103116605A (en) * | 2013-01-17 | 2013-05-22 | 上海交通大学 | Method and system of microblog hot events real-time detection based on detection subnet |
| CN103678474A (en) * | 2013-09-24 | 2014-03-26 | 浙江大学 | Method for acquiring large number of hot topics fast in social network |
Non-Patent Citations (1)
| Title |
|---|
| 刘威: "面向微博话题的用户影响力分析算法", 《计算机应用》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110209962A (en) * | 2019-06-12 | 2019-09-06 | 合肥工业大学 | The acquisition methods and system of theme level high-impact user |
| CN110209962B (en) * | 2019-06-12 | 2021-02-26 | 合肥工业大学 | Method and system for acquiring theme-level high-influence user |
| CN110910176A (en) * | 2019-11-27 | 2020-03-24 | 上海风秩科技有限公司 | Critical consumer recruitment method, apparatus, computer device and readable storage medium |
| CN110929172A (en) * | 2019-11-27 | 2020-03-27 | 中科曙光国际信息产业有限公司 | Information selection method and device, electronic equipment and readable storage medium |
| CN110929172B (en) * | 2019-11-27 | 2022-11-18 | 中科曙光国际信息产业有限公司 | Information selection method and device, electronic equipment and readable storage medium |
| CN111241420A (en) * | 2020-01-10 | 2020-06-05 | 云境商务智能研究院南京有限公司 | Recommendation method based on social network information diffusion perception |
| CN113781250A (en) * | 2020-09-14 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Social media information propagation evaluation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109800351A (en) | High-impact usage mining method in microblogging specific topics | |
| US11100411B2 (en) | Predicting influence in social networks | |
| Agarwal et al. | A model of crowd enabled organization: Theory and methods for understanding the role of twitter in the occupy protests | |
| Leskovec | Dynamics of large networks | |
| Centola | The social origins of networks and diffusion | |
| Budak et al. | Structural trend analysis for online social networks | |
| Chen et al. | An analysis of verifications in microblogging social networks--Sina Weibo | |
| Weng | Information diffusion on online social networks | |
| Yusriyah et al. | Communication networks analysis on information dissemination of the moving of capital city from Jakarta to East Kalimantan | |
| Borzymek et al. | Trust and distrust prediction in social network with combined graphical and review-based attributes | |
| Gadek et al. | Topical cohesion of communities on Twitter | |
| Kaligotla et al. | Diffusion of competing rumours on social media | |
| Shi et al. | Event detection and multi-source propagation for online social network management | |
| Bródka | A method for group extraction and analysis in multilayer social networks | |
| Liu et al. | Analysis of and defense against crowd-retweeting based spam in social networks | |
| Bargar et al. | Challenges and opportunities to counter information operations through social network analysis and theory | |
| Wang et al. | Predicting poi visits with a heterogeneous information network | |
| De Meo et al. | Improving the compactness in social network thematic groups by exploiting a multi-dimensional user-to-group matching algorithm | |
| Hajeer et al. | Clustering online social network communities using genetic algorithms | |
| Chiu et al. | Propagating online social networks: via different kinds of weak ties | |
| Dhakal et al. | Predicting friendship strength for privacy preserving: a case study on Facebook | |
| Kashyap et al. | LogRank: Summarizing Social Activity Logs. | |
| Vrana et al. | A network analysis of Greek tech blogs: A lonely road | |
| Rungsawang et al. | Un-biasing the link farm effect in pagerank computation | |
| Reelfs | Content & user behavior in anonymous hyperlocal online platforms |
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: 20190524 |
|
| RJ01 | Rejection of invention patent application after publication |