165 results about "Community finding" patented technology

The invention discloses an information recommending method based on a social network. The information recommending method includes the following steps that first, trust degree and similarity between users are calculated, and a user relation matrix is constructed through weighted values; second, the users are clustered through a community discovering algorithm, and then a closest neighbor set of the users is formed; third, scores are predicted, and a recommending list is generated. The information recommending method based on the social network can achieve the following advantages that first, the cold start problem is solved: trust degree is introduced into the method, if enough neighbors cannot be obtained according to the common grading articles in the recommending process, trustable friends can serve as the start point of prediction, and thus the cold start problem can be relieved, and user coverage can be improved; real time performance is improved: community division is performed on the user network through the community discovering algorithm commonly used in social network analysis, in other words, same user interests are clustered, and thus the time for finding the neighbor set of the users is greatly shortened, and the real time performance of the information recommending response is improved.

The invention discloses a financial network unusual transaction community finding method based on information entropy. The method includes the specific steps of firstly, analyzing the process and the mode of an unusual transaction, and extracting features of block transactions and suspicious transactions; secondly, storing transaction amounts and traction frequency information of accounts through an adjacent matrix so as to quantize the features; thirdly, defining the information entropy of nodes; fourthly, conducting community division and finding on the financial network nodes according to the financial network unusual transaction community finding method based on the information entropy; fifthly, evaluating the community finding result. Firstly, accuracy of unusual transaction account identification and unusual transaction communication division which are achieved according to the method is calculated, and then the structure of the unusual transaction community is evaluated through the community evaluation index modularity Q. The community finding method based on the information entropy is applied to the anti-money laundering field for the first time, the money laundering accounts can be identified and tracked by finding the unusual transaction community, and unusual transactions can be prevented by analyzing the relations and the transaction laws between money laundering accounts.

The invention discloses a group and place recommendation method based on locations and social relationships. The group and place recommendation method comprises the steps of: acquiring user check-in information in an LBSN, removing places and user data with poor effectiveness, and finally acquiring check-in data of users; utilizing a Pearson correlation coefficient, measuring check-in similarity based on common check-in data of the users, calculating a check-in similar degree among the users, and establishing a user check-in similar degree network; identifying different communities by utilizing a discrete particle swarm optimization method according to the user check-in similar degree network; and acquiring a friend list according to accounts of a user social network, forming social adjacent relationships of users in communities, finally generating a social group, recommending the social group to target users and recommending places to the target users by adopting the collaborative filtering recommendation method. The community finding method is simple and easy to operate, and the community dividing speed is fast. By adopting the method of combining the social group with place recommendation, the complexity of the method is reduced, and the recommendation precision is improved.

The invention provides a community discovery method. The community discovery method includes the steps that community division is conducted on a plurality of nodes in a network based on modularity maximization, and community boundary nodes obtained in the last step are adjusted based on community property entropy minimality. The community discovery method further includes the steps that if the community division obtained after adjustment meets an end condition, the community division will be used as final community division; otherwise, the communities obtained after adjustment will be used as nodes, community division will be conducted on the nodes again, and the community boundary nodes need to be readjusted. According to the community discovery method, the structure of the network and the attributive characters of the nodes are taken into account at the same time, and the degree of accuracy of community discovery is improved. In addition, the community discovery method is close to be linear in time complexity and suitable for large-scale on-line social network data.

The invention discloses a social network establishment method and a device, and a community discovery method and a device, and the social network establishment method comprises the following steps: respectively extracting feature words from all information units, and calculating feature vectors which correspond to all the information units according to the feature words; respectively calculating the similarity between each two information units according to the feature vectors; and establishing a social network according to the calculated similarity between each two information units. The method and the device can more really reflect the links among nodes in the network, and better carry out community division on the weighted network.

The invention relates to a method and device for detecting large-scale social network communities. The method includes the steps that input large-scale social networks are modeled into a map G = (V, E); all nodes on the map G are sequenced in a descending order according to the size relation of node degrees through a parallel sorting algorithm, and the sum DSum of effective degrees of all the nodes on the map G is calculated; DSum / P serves as an equally-dividing benchmark reference value, and the map G is equally divided into P sub maps through a load balancing method; the P sub maps are traversed for looking for triangles on the map G through a MapReduce parallel computing model, parallel multilayer coarsening is conducted on the map G based on a triangle topological structure, and then a simplest coarsened reduction map G' is obtained; by the adoption of a community finding algorithm based on genetics, initial community finding is conducted on the simplest coarsened reduction map G', and then a community finding result is generated; the community finding result is coarsened reversely layer by layer and restored to the map G, fine adjustment and optimizing processing is conducted accordingly, and then a community structure of the map G is acquired. According to the method and device, community finding and data analysis of the large-scale social networks can be accurately and efficiently achieved.

The invention discloses a community discovering method which comprises the following steps of: analyzing the information of each user, extracting a feature word from the information, and calculating a feature vector corresponding to the user; calculating the similarity of other users with the user in a mode that one user is taken as the reference; enabling a user with the similarity higher than a threshold as a similar user, combining all similar users as a new user, recording the combined similar users as a sub-user of a new user, and calculating and simplifying the feature vector of the new user; and accomplishing the discovery of new communities till the new user achieves a set community discovery threshold. According to the method, the community division is carried out according to the multi-attribute similarity of users without depending on the community behaviors of the users, users with the similar interests, research directions and behavior manners can be organized in one community once a social network is formed, and the discovered community is rich in attribute, the similarity of community users is high, and a relatively ideal platform for information communication and sharing can be provided for users.

Disclosed is an overlapping community discovery method for a network. Network nodes are initialized through non-overlapping community division results, and overlapping points among communities are solved through a fuzzy cooperative game based method to achieve overlapping community division of the network. The method comprises the steps of step 1, selecting a non-overlapping community discovery method for community division, wherein community overlapping points are not included; step 2, re-determining a membership function and calculating the degree of membership of the nodes to each community; step 3, re-determining a revenue function and calculating community revenues after nodes are added in the communities; step 4, solving the overlapping points of the communities according to the calculated degree of membership and the community revenues after the nodes are added in the communities to achieve network overlapping community division. According to the overlapping community discovery method, the network overlapping community discovery problem is solved through a novel idea, and the accuracy of the network overlapping community discovery method is improved.

The invention discloses a method for detecting a 'sockpuppet' based on a 'similar view' network and an article author judging technique. The method comprises the three steps that first, the 'similar view' network is established according to the mutual information of virtual social users; then the 'similar view' network is cut according to the writing styles of the users; at last, a community discovery algorithm is used for conducting community dividing on the cut 'sockpuppet' network, and the IDs in the same community are considered as the 'sockpuppet' of a certain person. The method for detecting the 'sockpuppet' has the following advantages of following the actual significance of the 'sockpuppet' community, being capable of being applied to detecting the 'sockpuppet' under a real-time network environment, and increasing the effectiveness of community discovery. The method for detecting the 'sockpuppet' is mainly applied to the various fields of public opinion analysis in virtual space, 'sockpuppet' detection and the like.

The invention discloses a community discovery method based on a Louvain algorithm. The method comprises the following steps that: S1: initializing a community, and taking each node as a community; S2:distributing each node to a community where each neighbour node is positioned in sequence to construct a community graph; S3: according to the community graph, taking the community as a node, and re-constructing the community graph; and S4: repeating S3, and outputting a result until all states are stable. The invention also discloses a community discovery system based on the Louvain algorithm. By use of the method and the device, each node in the network is taken as the community, and analysis is carried out by aiming at the modularity and the edge weight of the community so as to obtain more accurate community discovery.

The invention provides a controlled host detection method and device based on a knowledge graph, and the method comprises the steps: filtering extracted feature data by adopting data, in a response state of NXDOMAIN, in DNS flow, describing the data based on a knowledge graph construction framework, constructing an NXDOMAINIP knowledge graph; and finally, analyzing the knowledge graph by utilizinga community discovery algorithm and a community judgment algorithm to obtain a controlled host list and mark a suspected malicious code family. Based on a distributed data flow processing framework,real-time DNS request flow is analyzed on a large scale, and the analysis efficiency is greatly improved through a multi-stage data preprocessing process; and through NXDOMAINIP knowledge graph construction, community discovery and community judgment, word list splicing DGA domain names are detected, controlled hosts and related malicious code families are determined, victims are reminded in timeto carry out AV upgrading and full-disk scanning on the hosts, and host vulnerabilities are reinforced.

The invention puts forward a community discovery method used for a complex network. An algorithm designs a node transition probability through the analysis of a network topology structure, the significance of the node for the network community is evaluated on the basis of a random walk method, then, the node with the high significance is taken as a core to construct the network community, and finally, a community structure is regulated through a community edge trimming method. Compared with an existing method based on random walk, CDATP (Community Detection Algorithm Based on Asymmetric Transfer Probability of Nodes) exhibits the node design transition probability in the network, and the degree of importance of the node for the community is evaluated through the local transition of the node.

The invention provides a community discovery method which comprises the following steps: delimiting a search region in the scale range of a to-be-discovered community in a social network; pruning in the search region according to the number of neighbor nodes of a node, pruning a node with a neighbor node number less than the closeness of the to-be-discovered community from the social network; selecting one node in the rest nodes in the pruned social network, searching a community with the size of absolute value of S-1 in the neighbor node of the node, then forming the to-be-discovered community by combining the node and the searched community with the size of absolute value of S-1, adding the to-be-discovered community in a result set, wherein the absolute value of S represents the size of the community which is desired to discover; and moving the left margin of the searched region to left, and then repeatedly executing the previous steps in an enlarged search region until the search region achieves the minimum value of the scale of the to-be-discovered community.

The invention relates to a distributed recommendation method based on a Spark platform. When related parameters input by users are legal and historical behavior data clicked by the users is not empty, a recommendation sequence A based on ItemBased collaborative filtering is generated; community finding is carried out with the users as the vertexes and the number of common clicks of the users as the edge, and a recommendation sequence B based on the UserBased collaborative filtering is generated; the A and the B are merged according to different weights to generate a recommended sequence C based on collaborative filtering; on the basis of the C, personal click historical behaviors of the users are paid attention to, training is carried out by utilizing a factor decomposition machine model, a training model is generated for prediction, and a prediction recommendation sequence result P is generated; the C and the P are merged according to a merging rule, a final recommendation sequence F is generated and ordered, and the final recommendation sequence F is written into a real-time database. By means of the method, the requirement for recommending massive big data can be met, and collective wisdom recommendation and personal wisdom recommendation are combined to form the final recommendation sequence.

The invention provides a multi-resolution community discovering method based on fuzzy clustering. According to local interaction information of adjacent nodes, the structural similarity is introduced for measuring the fuzzy relation between the nodes, fuzzy transitivity of the fuzzy similarity between the nodes in a network topology is partially considered, fuzzy parameters are used for performing set cutting on a fuzzy transitivity matrix to obtain community structures under different resolutions, and therefore network communities can be discovered. Matrix transformation operation is adopted, a network community detection model based on fuzzy clustering is built, iterative optimization processes in a traditional method are reduced, the time complexity is lowered, a large number of experiments prove that the community structures in a network can be effectively revealed, the universality is strong, and the high application value is achieved; network structural analysis and community structural visualization can be effectively achieved.

The invention provides an enhanced network representation learning method based on community perception and relationship attention, and the method comprises the following steps: firstly obtaining a network topology structure and the text information of a node; secondly, acquiring community information by adopting a community discovery algorithm, marking a community where each node is located, andcombining the community with the topological structure to generate a community structure network; then, learning structure embedding of nodes on the community structure network by adopting a communityperception module; then, learning text embedding of each pair of adjacent nodes by adopting a relation attention module; and finally, performing model training in combination with structure embeddingand text embedding to obtain embedded representation of each node. According to the invention, local and global structures of the network and rich semantic relationships among nodes can be completelycaptured.

The present invention relates to a microblog big data interest community analysis optimization method based on a user experience. The method comprises a step of carrying out weighted reconfiguration of an original microblog network, a step of completing the community division of the reconfigured weighted network based on the discovery algorithm of a link community, and a step of using a hierarchical clustering algorithm to continuously merge two communities with a largest similarity, finally forming a link community through division, and generating a tree-shaped hierarchical diagram. Starting from aspects of interest modeling and community discovery, through analysing microblog content and a user behavior, a user is helped to find interested users and topics of the user. Compared with a traditional method, the accuracy, recall ratio and F value of the method of the invention are improved significantly.

The invention discloses a method and a device for identifying an associated specific account, and a storage medium. The method comprises the following steps: firstly, constructing an isomorphic directed graph according to associated data between accounts within a period of time; then, according to some specific modes of the specific account, finding out a seed node which may be the specific account from the isomorphic directed graph; then, finding a connected sub-graph containing the seed nodes; dividing the connected sub-graph into a plurality of communities according to a community discoveryalgorithm, and evaluating each community to obtain a score corresponding to each community; and then, determining the community with the relatively high risk as the target community, and wherein theaccount represented by the node in the target community is the identified associated specific account. Therefore, by using the method, associated specific accounts can be automatically identified fromlarge-scale associated data by using data mining methods such as graph theory, graph operation, community discovery, social network analysis and the like through unsupervised learning.

The invention discloses a community discovery algorithm in an enterprise map based on relationship weights. The community discovery algorithm comprises the following steps: (1) weights of relationships in the enterprise map are determined; (2) the similarity between enterprises are determined; (3) Modularity is used to evaluate the results of community discovery in complex networks; (4) convergingis carried out based on modularity constraint algorithm ; (5) Confidence probability table is calculated based on LPA algorithm. By discovering the communities in the enterprise map, the algorithm divides the entire enterprise map into sub-graphs. The divided enterprise map can not only help to analyze the industry prospects according to its community attributes, but also can be used to analyze the financial guarantee information, cooperation information, investment information and so on. And according to community attributes, the government or relevant departments are helped to formulate policies to adjust the market in a timely manner.

The invention relates to a microblog data management system and an implementation method thereof, and provides a service of managing microblog data through automatic friend grouping for a microblog user. The system consists of five modules, i.e. a user authorization module, a data extraction module, a community structure finding module, a grouping analysis and exhibition module, a feedback module and a microblog data management module. The system and the method have the advantages that the problems of waste at time and labor and difficult maintenance of traditional manual microblog data management are solved; the friends of the user are intelligently grouped by a community finding technique, so the accuracy is high, the overlapped communities can be found, and the like; a result is analyzed by the method to provide the visual and easy-understanding user friend grouping basis; in addition, the system provides a feedback mechanism to further improve the reliability of the system through introducing the feedback of the user into the system.

The invention relates to a label propagation community discovery algorithm based on node importance, and its main technical features are: initializing the unique label of each node; calculating the importance of each node, and sorting the nodes according to the node importance from high to low, Generate an ordered sequence; set the number of iterations t=1; for any node in the ordered sequence, update the label of the node to the label with the greatest influence in the label set of adjacent nodes according to the label selection method and label update rule; if the number of iterations t==max Iter or the label of each node is the most influential label, then the nodes with the same label are classified into the same community, and the process ends; otherwise, the number of iterations t is increased by 1, and the update is continued. The present invention has reasonable design, can significantly improve the quality of community discovery under the condition of similar complexity, shorten the iteration cycle, has high accuracy and stability, and can be widely used in community discovery, social network and other fields.

The invention relates to the technical field of data processing, and provides a community discovery method and device based on entity similarity in a knowledge graph. The method comprises the following steps: storing a social network data by using the knowledge graph, and calculating a Jacquard distance to obtain a similarity matrix; calculating a similar node set in the knowledge graph accordingto the similarity matrix; and performing iterative label propagation according to the similar node set, and determining a final community label of the node according to a label list of each node afterthe iteration to perform community discovery. A community network is stored by using the knowledge graph, thereby avoiding the storage of structure of missing data, and on the basis, the Jacquard distance is used as the basis for calculating the similarity, so that the accuracy is higher.

The invention discloses a coverable clustering algorithm applying to community discovery. The coverable clustering algorithm includes firstly, converting acquired original data into 'user-attribute graphs', primarily classifying behaviors in the 'user-attribute graphs' after initializing candidate subgraphs; secondly, computing domination attribute of each candidate subgraph while computing correlation between each user and each candidate subgraph; thirdly, establishing a probability statistical model, computing correlation between each 'user-attribute' pair and the candidate subgraph, iteratively constructing the candidate subgraphs until forming stable and effective candidate subgraph structures; and finally, reasonably classifying the different 'user-attribute' pairs in data according to the constructed candidate subgraphs in the data environment, and discovering key users having various attributes. The coverable clustering algorithm is used for processing content data and relative data simultaneously, and meets requirements on community discovery in real network environments well.

The invention discloses a community discovery method based on parallelization modularity optimization. The community discovery method comprises the following steps: initializing a network including N nodes into N communities, namely each node is an independent community, and initializing the modularity gain of a community pair in the local area world of each community; adopting a parallelization processing mechanism, searching community pairs with maximum modularity gains in the local area world of each community, merging the community pairs into the same community, and updating the modularity gains of the community pairs connected with new communities; repeating the step two until the community structure in the network is not changed. Community division is carried out on the network with greedy algorithm, and the processing time is extremely short by using parallelization manner and high-performance data structure. The method is suitable for the fields of automatic detection of large-scale dense network community structure, friend recommended system of social networking services and the like.

The invention discloses an online social network (short for OSN) community discovery method based on a Latent Dirichlet Allocation (short for LDA) theme model. The method comprises the following steps first pre-processing data, building an LDA theme model (including an LDA-F model and an LDA-T model) based on a relationship between a user in the online social network and other friends and word information expressed by the user to solve a model probability distribution, then estimating parameters via a Gibbs sampling algorithm, and at last discovering an OSN community according to the estimated parameters. By the use of the OSN community discovery method based on the LDA Theme model, a corresponding probability model can be achieved based on user blog semantic information discovery without the use of information connection via the network topology; blog content semantic similarities are introduced to effectively describe user interest and hobby probability distribution conditions; and with the introduction of community internal topological connection tightness, communities with close internal topological connections can be discovered.

The invention discloses a method for recommending digital television programs based on community finding. The method comprises a step 202 of obtaining data of a multilayer social network relevant to users by a program recommendation system according to login information of the television users, a step 204 of carrying out community division on the multilayer social network by the program recommendation system through the community finding method, and a step 206 of sending corresponding program recommending information to a digital television receiving end of each community member with each community as a unit by the program recommendation system according to community division situations obtained in the step 204. The method for recommending the digital television programs utilizes information contained in a social network structure to recommend content of the television programs with each community as the unit, so that comprehensiveness and accuracy of a recommendation result are improved.

The invention relates to a community discovery method based on a directed graph of a social network. According to the invention, based on different types of triangles, the features of triangles are extracted and quantified as edge weights between points. After that, a directed unauthorized graph is directly converted into a non-directed unauthorized graph. After that, the community discovery is realized by adopting the improved label propagation algorithm. According to the method, the problem that the traditional community discovery algorithm in the social network does not apply to the directed graph in the prior art can be solved. Meanwhile, the accuracy of the community division is improved greatly through the improvement of the algorithm.

The invention relates to the community discovery field in a complex network, and specifically discloses a community discovery method based on local optimization. On the basis of an LFM (Local FitnessMaximum) algorithm, on the one hand, a weighting method for integrating internodal boundary social attribute with common neighbor node degree in the network is proposed, and a fitness function in theLFM algorithm is updated by using the formed boundary weight; on the other hand, the concept about the local community stability is imported into the local optimization process of the LFM algorithm, the contribution degree on the community stability by the node is judged by computing the community stability before and after the node adds in the local community, and the contribution degree is usedas the criterion whether the node is added in the community. Through the method disclosed by the invention, the appearance of the over-sized community is avoided, the more significant groupuscule structure in the network can be easily discovered, and the method is suitable for the real social network.

At present, an existing social network community quick division algorithm has the problems of low quality and incapability of fully utilizing node link information, and a division algorithm with a relatively good effect also has the problems of high time complexity and incapability of application to large-scale social networks. For the problems, a trust correlation based microblog network community discovery method is provided. On the basis of defining community node pair information group degree and dynamically distributing network edge weight values, a trust correlation matrix of nodes is calculated, the nodes are subjected to cluster analysis through an improved K-medoids algorithm, and an ideal structure of a network community is determined by calculating LC modularity of community quantity. An experiment is performed on a sina microblog data set, and a result shows that the algorithm can enable a community division result to be more accurate.