CN106780073B - Social network influence maximization initial node selection method considering user behaviors and emotions - Google Patents
Social network influence maximization initial node selection method considering user behaviors and emotions Download PDFInfo
- Publication number
- CN106780073B CN106780073B CN201710019177.XA CN201710019177A CN106780073B CN 106780073 B CN106780073 B CN 106780073B CN 201710019177 A CN201710019177 A CN 201710019177A CN 106780073 B CN106780073 B CN 106780073B
- Authority
- CN
- China
- Prior art keywords
- emotion
- influence
- nodes
- evaluation
- comment
- 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
- 230000008451 emotion Effects 0.000 title claims abstract description 100
- 230000006399 behavior Effects 0.000 title claims abstract description 60
- 238000010187 selection method Methods 0.000 title claims abstract description 11
- 230000002996 emotional effect Effects 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 15
- 238000011156 evaluation Methods 0.000 claims description 46
- 230000011218 segmentation Effects 0.000 claims description 21
- 230000008901 benefit Effects 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 11
- 230000009193 crawling Effects 0.000 claims description 4
- 230000001934 delay Effects 0.000 claims description 4
- 230000003542 behavioural effect Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000013075 data extraction Methods 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 102100035345 Cerebral dopamine neurotrophic factor Human genes 0.000 description 10
- 101000737775 Homo sapiens Cerebral dopamine neurotrophic factor Proteins 0.000 description 10
- 230000000694 effects Effects 0.000 description 9
- 238000002474 experimental method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000004941 influx Effects 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 231100000656 threshold model Toxicity 0.000 description 2
- 241000700605 Viruses Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a social network influence maximization initial node selection method considering user behaviors and emotions. The method comprehensively considers the user behavior tendency and the emotional tendency, and more effectively, accurately and truly excavates the initial node with maximized influence.
Description
Technical Field
The invention belongs to the technical field of computers, and relates to a social network influence maximization initial node selection method considering user behaviors and emotions.
Background
With the rapid development of social networks, more and more people express personal opinions and opinions of the people on other people, products, social events and the like through social networks such as microblogs, circle of friends, Facebook, Twitter and the like, share and communicate with other people, and show work, life and emotion of the people to the social networks. In the social network, the information of communication interaction between users shows the emotional tendency between users, and generally, the emotion between users is directly related to the influence factor between users. According to the emotional influence relationship of users in the social network, a user group with the largest influence is searched, so that the information is spread in a virus mode at the minimum cost, and the influence is spread in a public praise effect mode. Therefore, emotion analysis on social network users is of great significance to influence maximization research.
Many improved algorithms and models based on IC models (independent cascade models) and LT models (linear threshold models) are proposed to find the initial node with maximized influence. User emotion analysis and time delay-based user behavior trends in social networks are rarely considered when affecting power propagation.
Therefore, it is necessary to provide a social network influence maximization initial node selection method considering user behaviors and emotions.
Disclosure of Invention
The invention solves the technical problem that aiming at the defects of the prior art, the invention provides the method for selecting the maximum social network influence initial node considering the user behaviors and the emotions, the maximum marginal profit node is solved as the influence maximum initial node based on the behavior tendency and the emotion tendency of the users in the social network, and the result is more accurate and reliable.
The technical scheme of the invention is as follows:
a social network influence maximization initial node selection method considering user behaviors and emotions comprises the following steps:
step 1: constructing an emotion word training model by using an evaluation data set with classification labels, and obtaining an emotion word vocabulary with emotion weight suitable for network comment information emotion analysis on the basis of an existing emotion dictionary and expanding the emotion word vocabulary to obtain a new emotion dictionary;
because the evaluation information of the social network is label-free, the commodity evaluation information in the e-commerce network is adopted, the evaluation information is subjected to emotion classification according to the evaluation star level, wherein 5 stars are positive evaluation, and 1 star is negative evaluation; in the method, the crawler is used for crawling the evaluation information of the commodities in the Jingdong, positive and negative samples of the commodity evaluation are extracted, and the quantity balance of the positive and negative samples is kept. The method for constructing the model by evaluating the information emotion vocabulary comprises the following steps:
step 1.1, performing Chinese word segmentation on a commodity evaluation text by adopting ANSJ to obtain word segmentation results marked with word parts of speech, and counting the times of appearance of adjective word segmentation in a positive evaluation sample and a negative evaluation sample respectively;
step 1.2, counting the sum N of the number of positive evaluation samples and negative evaluation samples, and setting a word segmentation frequency threshold r and a maximum emotion weight max;
step 1.3, for adjective word, counting the occurrence frequency N of the adjective word in a forward evaluation samplePAnd the number of occurrences N in the negative evaluation sampleNIf, ifExecute1.4, 1.5 and 1.6;
step 1.4 calculating the forward probability P of the word segmentation wordP:
Wherein P isP∈(0,1);
Step 1.5, calculating the emotion weight w of the word according to the forward probability of the word1And w2:
Wherein P isPE (0,1), then w1∈(-1,1);
Wherein w2∈(-1,1);
Step 1.6, giving an adjective word, wherein the comprehensive emotion weight is w:
step 2: carrying out data preprocessing and data extraction on the social network data set, extracting user behavior records, and calculating user behavior influence considering user behavior time delay;
in the social network, if user nodes u and v are adjacent nodes, and v generates a series of behaviors for a message issued by u through a series of time delays, wherein the behaviors comprise praise, comment and forward, a user behavior influence B _ Inf (u, v) considering the user behavior time delays is as follows:
wherein λ ∈ (0,1), m (u) represents a message set of u executed behavior, m (v) represents a message set of user node v executed behavior, | m (u) | represents the number of message pieces of u executed behavior, | m (u) | n m (v) | represents the number of pieces of v executed behavior on u published message, T represents the average time delay of v on u published message execution behavior, and T represents the average time delay of v on u published message execution behavior.
And step 3: according to the expanded emotion dictionary, emotion analysis is carried out on comment information among social network users, and user emotion influence based on user emotion tendencies is obtained;
giving two adjacent user nodes u and v, if u issues a message, v evaluates the message, and evaluates one piece of evaluation content commentjPerforming word segmentation to obtain a series of word segmentation sets1,word2,...,wordnFinding corresponding participle emotion weight { w ] of each participle in the emotion dictionary1,w2,...,wnGet it beforewiIf 0, the emotion score of the piece of evaluation content:
words with turning meanings often appear in comment information, and the combination of the words and other emotion participles can reverse the meaning expressed by the emotion participles modified by the words; therefore, considering the modified adverb of the inflected semanteme, the context of the words is analyzed, and the emotion participles adjacent to the words are extracted to obtain the inflected participle set { reverse _ w1,reverse_w2,...,reverse_wmLooking up their respective emotion weights in emotion dictionary { r _ w }1,r_w2,...,r_wmTo commentjThe emotion value is corrected to obtain a corrected emotion value:
therefore, if the content comment is evaluatedjIs composed ofThe more the emotional words are, the higher the accumulated value of the segmentation emotion weight is, and the greater the positive emotion of v to u is, so that the influence of u to v can be deduced to be greater. If the content comment is evaluatedjThe more negative emotion words are contained in the segmentation emotion weight, the more negative and negative segmentation emotion weight accumulated values are, the stronger the negative emotion of u is represented by v, and therefore the larger the influence of u on v is deduced. Thus, given two adjacent user nodes u and v, v a comment for ujBased on commentjDeduces the individual emotional influence of u on v:
wherein ∈ (0, 1);
thus given two neighboring user nodes u and v and multiple pieces of comment information { comment of v on u1,comment2,...,commentnAnd (6) calculating the node emotional influence force according to each comment, and averaging to obtain the emotional influence force of u on v:
and 4, step 4: comprehensively considering the influence of emotional tendency based on user behavior tendency to obtain total influence TotalInf of user nodeu,v:
TotalInfu,v=β·B_Inf(u,v)+(1-β)·SentiInfu,v;
And 5: TotalInf the total influence of the user nodes calculated according to the step 4u,vAnd the topological relation of the social network nodes is comprehensively considered, so that the total propagation influence of u on v is phiu,v(v):
Wherein I (v) represents a set of nodes in the network topology that have behavioral and emotional impact on v;
step 6: calculating the total influence marginal gain of the node u as sigmaMarginal_Revenue(u):
The method comprises the following steps that S represents an initial node set, and A represents a set of all user nodes in a social network;
calculating the marginal benefits of the nodes in the social network, sequencing the marginal benefits according to the sizes, selecting the first k nodes with the maximum marginal benefits, sequentially inserting the first k nodes into a queue Q, popping the nodes with the maximum marginal benefits in the Q, and inserting the nodes into S, wherein k is the number of the initial node sets;
and 7: counting the number of nodes in S, if | S | < k, recalculating all the marginal benefits of the nodes in the queue Q according to step 6, updating and sequencing, popping the node with the maximum marginal benefit in Q, inserting the node into S, and if | S | ═ k, taking S as a final initial node set;
has the advantages that:
the invention provides a social network influence maximization initial node selection method considering user behaviors and emotions. Based on user Behaviors and emotional tendency in the online social network, an Influence propagation model comprehensively considering two factors is provided and named as BSIS (influx Spread Based on Users' Behaviors and Sentiment) model. The model is divided into three stages, wherein an emotion word training model expansion emotion dictionary is established in the first stage, user behavior influence is established in the second stage, and user emotion influence is established in the third stage. And solving the maximum marginal profit node by combining a greedy algorithm and a BSIS model to serve as the initial node with the maximum influence. And (3) crawling commodity evaluation data from the Kyoto, dividing the evaluation into a positive evaluation and a negative evaluation according to the evaluation star level, and expanding the existing emotion dictionary based on the emotion word training model. The model and the algorithm provided by the invention are realized on Hadoop and Spark platforms by using Flickr and microblog social network data sets. Experimental results show that compared with a method based on traditional IC, LT and CDNF models, the influence propagation model and the influence propagation algorithm provided by the invention have the advantages of better influence propagation effect, wider influence propagation range and higher initial node selection quality.
Drawings
FIG. 1 is a flow chart of a social network influence maximization initial node selection method considering user behaviors and emotions, which is provided by the invention;
FIG. 2 is a graph comparing the effect of influence propagation of initial nodes selected in example 1 by using Flickr data set based on 4 different methods of BSIS, CDNF, IC and LT;
fig. 3 is a graph of an influence propagation effect generated by an initial node selected based on BSIS using a microblog dataset in embodiment 1;
FIG. 4 is a comparison graph of the execution times of received praise, forward and comment behaviors of the initial node selected by 4 different methods of BSIS, CDNF, IC and LT using the Flickr dataset in the embodiment 1;
fig. 5 is a diagram of the number of times of performance of received praise, forward, and comment actions of an initial node selected based on BSIS using a microblog dataset in embodiment 1;
Detailed Description
In order to describe the social network influence maximization initial node selection method considering user behaviors and emotions, which is proposed by the invention, in more detail, the invention is further described with reference to the accompanying drawings and examples.
The invention provides a social network influence maximization initial node selection method considering user behaviors and emotions. Based on user Behaviors and emotional tendency in the online social network, an Influence propagation model comprehensively considering the two factors is provided and named as BSIS (influx Spread Based on Users' Behaviors and Sentiment) model. The model is divided into three stages, wherein an emotion word training model expansion emotion dictionary is established in the first stage, user behavior influence is established in the second stage, and user emotion influence is established in the third stage. And solving the maximum marginal profit node by combining a greedy algorithm and a BSIS model to serve as the initial node with the maximum influence. And (3) crawling commodity evaluation data from the Jingdong website, dividing evaluation into positive evaluation and negative evaluation according to evaluation star grades, and expanding the existing emotion dictionary based on an emotion word training model. The model and the algorithm provided by the invention are realized on Hadoop and Spark platforms by using Flickr and microblog social network data sets. Experimental results show that compared with the method on the traditional IC, LT and CDNF models, the influence propagation model and the influence propagation algorithm provided by the invention have the advantages of better influence propagation effect, wider influence propagation range and higher initial node selection quality.
Fig. 1 is a flowchart of a method for selecting an initial node maximizing social network influence in consideration of user behaviors and emotions, which is provided by the invention, and the method comprises the following specific implementation steps:
example 1:
in this example, crawler crawls the kyoto shopping web user comment data, which contains 17052 5-star positive ratings and 14937 1-star negative ratings, as a training set and test set of emotion word training models. In the embodiment, an emotion word training model is constructed based on the step 1, 80% is used as a training set, 20% is used as a test set, and finally 387 emotion words with weights are obtained through training to expand the existing Chinese emotion word list NTUSD and SentiWordNet English emotion word list of Taiwan university.
In this embodiment, a Flickr (one Picture sharing social networking site in the United states) social networking dataset is utilized, which contains 40808 user nodes and 75269 user behavior records. A microblog social network data set is also used, which contains 412952 user nodes and 500977 user behavior records.
The embodiment is realized on Hadoop2.5.2 and spark2.0.2(12Core, 120GB Memory, 6Workers), the BSIS influence propagation model and the initial node solving algorithm are effectively and quickly realized, and the influence maximization initial node set is obtained through calculation. A comparison experiment is designed from two aspects of influence propagation effect and user real influence range by utilizing two real social network data sets, and the correctness and the effectiveness of the method are verified by comparing and comparing our results with a traditional IC model (an independent cascade model), an LT model (a linear threshold model) and a CDNF model (a credit distribution model based on node characteristics). The edge activation probability between adjacent points in the IC model is learned through an EM (maximum expectation algorithm), and the edge activation probability pp (u, v) of adjacent nodes v and u of the LT model is calculated by using the formula pp (u, v) ═ 1/N (u), wherein N (u) represents the number of adjacent nodes of the node u.
As can be seen from fig. 2, the left sub-graph represents a total influence distribution graph calculated from the first 50 initial nodes solved based on the four models of SBIS, CDNF, IC, and LT in the Flickr social network, and the right sub-graph represents a distribution graph in which the total influence sum of the initial nodes solved based on the four models of SBIS, CDNF, IC, and LT increases with the number of the initial nodes. As can be seen from fig. 3, the left and right subgraphs show a total influence sum increase graph of a total influence distribution graph of the first 50 initial nodes solved based on the SBIS model in the microblog social network. As shown in verification of FIGS. 2 and 3, the SBIS comprehensively considers the behavioral factors of the user emotion, and the selected initial node has a larger influence propagation effect.
In order to verify the real influence generated by the initial node selected by the SBIS model in the real social network, the embodiment utilizes the times of praise, forwarding and comment behaviors received by the user and executed by other users as the real influence measurement standard, and counts the times of executing the real behaviors received by the initial node set selected by the SBIS, CDNF, IC and LT-based four influence propagation models. As can be seen from fig. 4, the Flickr data set is used, the left sub-graph represents a distribution graph of the execution times of the praise, forwarding and comment behaviors received by the first 50 initial nodes selected based on SBIS, CDNF, IC and LT, and the right sub-graph represents an increase graph of the sum of the execution times of the praise, forwarding and comment behaviors received by the initial nodes selected based on SBIS, CDNF, IC and LT with the increase of the number of the initial nodes. As can be seen from fig. 5, the left and right subgraphs respectively represent a graph of the behavior execution times of like, forward and comment received by the first 50 initial nodes solved by using the microblog data sets and the BSIS model, and a graph of the sum of the times. As can be seen from the graph, the execution times of the behaviors received by the initial node selected based on the SBIS influence propagation model are higher than those of other models, so that compared with other models, the SBIS can effectively mine hot users, the selection quality of the initial node is guaranteed, a wider influence propagation range is generated in a real social network, and the influence propagation effect is reflected more accurately.
From the experiments, the influence maximization initial node is mined based on the user behavior and emotional tendency factors, the influence of the user can be reflected more truly, and the experiments show that the method has accuracy and reliability in the aspects of initial node selection quality, influence propagation effect and influence propagation range.
Claims (1)
1. A social network influence maximization initial node selection method considering user behaviors and emotions is characterized by comprising the following steps:
step 1: constructing an emotion word training model by using an evaluation data set with classification labels, and obtaining an emotion word vocabulary with emotion weight suitable for network comment information emotion analysis on the basis of an existing emotion dictionary and expanding the emotion word vocabulary to obtain a new emotion dictionary;
because the evaluation information of the social network is label-free, the commodity evaluation information in the e-commerce network is adopted, the evaluation information is subjected to emotion classification according to the evaluation star level, wherein 5 stars are positive evaluation, and 1 star is negative evaluation; crawling the evaluation information of the goods in the Jingdong by using a crawler, extracting positive and negative samples of the goods evaluation, and keeping the quantity balance of the positive and negative samples; the method for constructing the model by evaluating the information emotion vocabulary comprises the following steps:
step 1.1, performing Chinese word segmentation on a commodity evaluation text by adopting ANSJ to obtain word segmentation results marked with word parts of speech, and counting the times of appearance of adjective word segmentation in a positive evaluation sample and a negative evaluation sample respectively;
step 1.2, counting the sum N of the number of positive evaluation samples and negative evaluation samples, and setting a word segmentation frequency threshold r and a maximum emotion weight max;
step 1.3, for adjective word, counting the occurrence frequency N of the adjective word in a forward evaluation samplePAnd the number of occurrences N in the negative evaluation sampleNIf, ifExecuting steps 1.4, 1.5 and 1.6;
and (1).4 calculating the forward probability P of word segmentationP:
Wherein P isP∈(0,1);
Step 1.5, calculating the emotion weight w of the word according to the forward probability of the word1And w2:
Wherein P isPE (0,1), then w1∈(-1,1);
Wherein w2∈(-1,1);
Step 1.6, giving an adjective word, wherein the comprehensive emotion weight is w:
step 2: carrying out data preprocessing and data extraction on the social network data set, extracting user behavior records, and calculating user behavior influence considering user behavior time delay;
in the social network, if user nodes u and v are adjacent nodes, and v generates a series of behaviors for a message issued by u through a series of time delays, wherein the behaviors comprise praise, comment and forward, a user behavior influence B _ Inf (u, v) considering the user behavior time delays is as follows:
wherein, λ ∈ (0,1), m (u) represents a message set of u executed behavior, m (v) represents a message set of user node v executed behavior, | m (u) | represents the number of message pieces of u executed behavior, | m (u) | n m (v) | represents the number of pieces of v executed behavior on u published message, T represents the average time delay of v on u published message execution behavior, and T represents the average time delay of v on u published message execution behavior;
and step 3: according to the expanded emotion dictionary, emotion analysis is carried out on comment information among social network users, and user emotion influence based on user emotion tendencies is obtained;
giving two adjacent user nodes u and v, if u issues a message, v evaluates the message, and evaluates one piece of evaluation content commentjPerforming word segmentation to obtain a series of word segmentation sets1,word2,...,wordnFinding the corresponding participle emotion weight of each participle in an emotion dictionary, { w1,w2,...,wnGet it beforeEmotional dictionary, wiIf 0, the emotion score of the piece of evaluation content:
words with turning meanings often appear in comment information, and the combination of the words and other emotion participles can reverse the meaning expressed by the emotion participles modified by the words; therefore, considering the modified adverb of the inflected semanteme, the context of the words is analyzed, and the emotion participles adjacent to the words are extracted to obtain the inflected participle set { reverse _ w1,reverse_w2,...,reverse_wmLooking up their respective emotion weights in emotion dictionary { r _ w }1,r_w2,...,r_wmTo commentjThe emotion value is corrected to obtain a corrected emotion value:
therefore, if the content comment is evaluatedjThe more forward emotional words are contained in the word segmentation emotion weight, the higher the accumulated value of the word segmentation emotion weight is, the word segmentation emotion weight is positive, the stronger the forward emotion of u represented by v is, and therefore the larger the influence of u on v can be deduced; if the content comment is evaluatedjThe more negative emotion words are contained in the segmentation emotion weight, the more negative and negative segmentation emotion weight accumulated values are, the stronger the negative emotion of u is represented by v, and the larger the influence of u on v is deduced; thus, given two adjacent user nodes u and v, v a comment for ujBased on commentjDeduces the individual emotional influence of u on v:
wherein ∈ (0, 1);
thus given two neighboring user nodes u and v and multiple pieces of comment information { comment of v on u1,comment2,...,commentnAnd (6) calculating the node emotional influence force according to each comment, and averaging to obtain the emotional influence force of u on v:
and 4, step 4: comprehensively considering the influence of emotional tendency based on the behavior tendency of the user to obtain the total influence TotalInf of the useru,v:
TotalInfu,v=β·B_Inf(u,v)+(1-β)·SentiInfu,v;
And 5: the total influence propagation path of u to v is obtained by comprehensively considering the topological relation of the social network nodes, so that the total of the propagation influence of u to v is phiu,v(v):
Wherein I (v) represents a set of nodes in the network topology that have behavioral and emotional impact on v;
step 6: calculating the total influence marginal gain of the node u as sigmaMarginal_Revenue(u):
The method comprises the following steps that S represents an initial node set, and A represents a set of all user nodes in a social network;
calculating the marginal benefits of the nodes in the social network, sorting the marginal benefits according to sizes, selecting the first nk nodes with the largest marginal benefits, sequentially putting the nodes into a queue Q, popping up the nodes with the largest marginal benefits in the Q, and putting the nodes into an S;
and 7: counting the number of nodes in S, if | S | < k, recalculating all the marginal benefits of the nodes in the queue Q according to step 6, updating and sequencing, popping the node with the maximum marginal benefit in Q, and putting the node into S, and if | S | ═ k, then S is a final initial node set.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710019177.XA CN106780073B (en) | 2017-01-11 | 2017-01-11 | Social network influence maximization initial node selection method considering user behaviors and emotions |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710019177.XA CN106780073B (en) | 2017-01-11 | 2017-01-11 | Social network influence maximization initial node selection method considering user behaviors and emotions |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106780073A CN106780073A (en) | 2017-05-31 |
CN106780073B true CN106780073B (en) | 2021-05-25 |
Family
ID=58949265
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710019177.XA Active CN106780073B (en) | 2017-01-11 | 2017-01-11 | Social network influence maximization initial node selection method considering user behaviors and emotions |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106780073B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108460010A (en) * | 2018-01-17 | 2018-08-28 | 南京邮电大学 | A kind of comprehensive grade model implementation method based on sentiment analysis |
CN108549632B (en) * | 2018-04-03 | 2022-02-11 | 重庆邮电大学 | Social network influence propagation model construction method based on emotion analysis |
CN108804524B (en) * | 2018-04-27 | 2020-03-27 | 成都信息工程大学 | Emotion distinguishing and importance dividing method based on hierarchical classification system |
CN110598960B (en) * | 2018-05-23 | 2022-06-03 | 北京国双科技有限公司 | Entity-level emotion assessment method and device |
CN110138619B (en) * | 2019-05-28 | 2020-05-19 | 湖南大学 | Initial node selection method and system for realizing influence maximization |
CN110738421B (en) * | 2019-10-17 | 2023-08-22 | 西南大学 | Multilayer network user influence measuring method based on shortest propagation path |
CN110727881B (en) * | 2019-10-23 | 2022-08-09 | 北京秒针人工智能科技有限公司 | Method and device for determining target public figure |
CN112052995B (en) * | 2020-08-31 | 2023-08-01 | 杭州电子科技大学 | Social network user influence prediction method based on fusion emotion tendency theme |
CN112487304B (en) * | 2020-11-26 | 2022-05-03 | 杭州电子科技大学 | Method for establishing influence propagation model based on viewpoint vectorization |
CN113688202B (en) * | 2021-07-30 | 2024-03-15 | 杭州网易云音乐科技有限公司 | Emotion polarity analysis method and device, electronic equipment and computer storage medium |
CN116245555B (en) * | 2023-03-09 | 2023-12-08 | 张家口巧工匠科技服务有限公司 | User information collecting and analyzing system based on big data |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101770487A (en) * | 2008-12-26 | 2010-07-07 | 聚友空间网络技术有限公司 | Method and system for calculating user influence in social network |
CN103530360A (en) * | 2013-10-12 | 2014-01-22 | 广西师范学院 | Network society influence maximization algorithm based on microblog text affective computing |
-
2017
- 2017-01-11 CN CN201710019177.XA patent/CN106780073B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101770487A (en) * | 2008-12-26 | 2010-07-07 | 聚友空间网络技术有限公司 | Method and system for calculating user influence in social network |
CN103530360A (en) * | 2013-10-12 | 2014-01-22 | 广西师范学院 | Network society influence maximization algorithm based on microblog text affective computing |
Non-Patent Citations (1)
Title |
---|
Credit Distribution and Influence Maximization in;Xiaoheng Deng等;《2015 12th International Conference on Fuzzy Systems and Knowledge Discovery》;20151231;第2093-2100页 * |
Also Published As
Publication number | Publication date |
---|---|
CN106780073A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106780073B (en) | Social network influence maximization initial node selection method considering user behaviors and emotions | |
CN105183717B (en) | A kind of OSN user feeling analysis methods based on random forest and customer relationship | |
Peddinti et al. | Domain adaptation in sentiment analysis of twitter | |
CN106355506B (en) | Influence maximization initial node selection method in online social network | |
CN110009430B (en) | Cheating user detection method, electronic device and computer readable storage medium | |
Shirsat et al. | Document level sentiment analysis from news articles | |
Probierz et al. | Rapid detection of fake news based on machine learning methods | |
CN112199608A (en) | Social media rumor detection method based on network information propagation graph modeling | |
Siddharth et al. | Sentiment analysis on twitter data using machine learning algorithms in python | |
Lu et al. | Exploring the sentiment strength of user reviews | |
CN110110220A (en) | Merge the recommended models of social networks and user's evaluation | |
de Zarate et al. | Measuring controversy in social networks through nlp | |
Biswas et al. | Sentiment analysis on user reaction for online food delivery services using bert model | |
Modi et al. | Sentiment analysis of Twitter feeds using flask environment: A superior application of data analysis | |
Elbaghazaoui et al. | Data profiling and machine learning to identify influencers from social media platforms | |
CN104572623B (en) | A kind of efficient data analysis and summary method of online LDA models | |
CN114842247B (en) | Characteristic accumulation-based graph convolution network semi-supervised node classification method | |
de Sousa et al. | Social network advertising classification based on content categories | |
Ahmad et al. | Google maps data analysis of clothing brands in south punjab, pakistan | |
Raj et al. | Automated Cyberstalking Classification using Social Media | |
Contreras et al. | Lexicon-based Sentiment Analysis with Pattern Matching Application using Regular Expression in Automata | |
Patel et al. | Rumour detection using graph neural network and oversampling in benchmark Twitter dataset | |
Turdjai et al. | Simulation of marketplace customer satisfaction analysis based on machine learning algorithms | |
Narmadha et al. | Recognizing eminent players from the Indian Premier League using CNN model | |
Sagvekar et al. | Study on product opinion analysis for customer satisfaction on e-commerce websites |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |