WO2019042060A1 - 成员角色确定方法、装置及存储介质 - Google Patents

成员角色确定方法、装置及存储介质 Download PDF

Info

Publication number
WO2019042060A1
WO2019042060A1 PCT/CN2018/097759 CN2018097759W WO2019042060A1 WO 2019042060 A1 WO2019042060 A1 WO 2019042060A1 CN 2018097759 W CN2018097759 W CN 2018097759W WO 2019042060 A1 WO2019042060 A1 WO 2019042060A1
Authority
WO
WIPO (PCT)
Prior art keywords
group
members
communication
node
network
Prior art date
Application number
PCT/CN2018/097759
Other languages
English (en)
French (fr)
Inventor
李志颖
胡麟
Original Assignee
腾讯科技(深圳)有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 腾讯科技(深圳)有限公司 filed Critical 腾讯科技(深圳)有限公司
Publication of WO2019042060A1 publication Critical patent/WO2019042060A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35Clustering; Classification
    • G06F16/355Class or cluster creation or modification
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/958Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2216/00Indexing scheme relating to additional aspects of information retrieval not explicitly covered by G06F16/00 and subgroups
    • G06F2216/03Data mining

Definitions

  • TECHNICAL FIELD Embodiments of the present application relate to the field of data mining, and in particular, to a member role determining method, apparatus, and storage medium.
  • BACKGROUND OF THE INVENTION There are various social groups in real life, such as family groups, trading groups, etc. These social groups include multiple members, each member playing a different role, how to determine the role of members has become a society The most closely related problem of life.
  • a virtual network is generally created according to a member communication record of a specified community, the virtual network includes multiple nodes, each node represents one member of a specified community, and two of the two members communicated Node connections form an edge.
  • the hierarchical clustering algorithm is used to cluster the members in the virtual network to obtain multiple sub-networks included in the virtual network
  • the Spline (spline interpolation) algorithm is used to calculate the centrality of each node in each sub-network.
  • the centrality may represent the closeness of the self-centered sub-network controlled by the member, and determine the member corresponding to the node whose central degree is greater than the preset number of times as the core member of the designated group.
  • a method for determining a member role is provided, which is applied to an analysis device, and the method includes:
  • the virtual network includes a plurality of nodes, each node represents one member of the specified community, and two nodes corresponding to any two members communicated form a side ;
  • the core sub-network is a network that has a degree of removal less than a preset number of cores and a network that is less than a side of a node whose degree is less than a preset number of cores;
  • An adjacency matrix is created in the core sub-network, the adjacency matrix is a row and a column in a node in the core sub-network, and an arrangement order of the row nodes and an arrangement order of the column nodes are the same, and each position in the adjacency matrix
  • the upper element is determined according to the connection relationship between the corresponding row node and the column node;
  • the adjacency matrix Starting from the adjacency matrix, repeatedly calculating a correlation coefficient between each node in the first dimension until a specified matrix including only the first element and the second element is acquired, the first dimension being a row or a column, where the One element is larger than the second element;
  • a member role determining apparatus which is applied to an analyzing device, the apparatus comprising a processor and a memory, wherein the memory stores an instruction executable by the processor, when the instruction is executed,
  • the processor is used to:
  • Creating a virtual network according to a member communication record of a specified group the virtual network including a plurality of nodes, each node representing one member of the specified community, and any two communicated The two nodes corresponding to the member are connected to form an edge;
  • the core sub-network is a network that has a degree of removal less than a preset number of cores and a network that is less than a side of a node whose degree is less than a preset number of cores;
  • the adjacency matrix uses nodes in the core sub-network as rows and columns, and the row order of the row nodes and the column nodes are arranged in the same order, and each position in the adjacency matrix
  • the upper element is determined according to the connection relationship between the corresponding row node and the column node;
  • the adjacency matrix Starting from the adjacency matrix, repeatedly calculating a correlation coefficient between each node in the first dimension until a specified matrix including only the first element and the second element is acquired, the first dimension being a row or a column, where the One element is larger than the second element;
  • a third aspect provides a computer readable storage medium, where the computer readable storage medium stores at least one instruction, at least one program, a code set, or a set of instructions, the instruction, the program, and the code set Or the set of instructions is loaded and executed by a processor to implement the operations performed in the member role determining method as described in the first aspect.
  • FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the present application.
  • FIG. 2 is a flowchart of a method for determining a member role according to an embodiment of the present application
  • 3 is a schematic diagram of a core sub-network provided by an embodiment of the present application
  • 4A is a schematic diagram of a core sub-network provided by an embodiment of the present application.
  • 4B is a schematic diagram of adding a self-edge according to an embodiment of the present application.
  • 4C is a schematic diagram of an operation flow provided by an embodiment of the present application.
  • FIG. 5 is a schematic diagram of a member role determining method according to an embodiment of the present disclosure
  • FIG. 6 is a schematic structural diagram of a member role determining apparatus according to an embodiment of the present disclosure
  • FIG. 8 is a schematic structural diagram of a server according to an embodiment of the present application.
  • the technical solutions in the embodiments of the present application are clearly and completely described in the following with reference to the drawings in the embodiments of the present application. It is obvious that the described embodiments are a part of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present application without creative efforts are within the scope of the present application.
  • the implementation environment of the embodiment of the present application may include an analysis device, where the analysis device acquires description information of a specified community, where the description information includes multiple members in a specified community and a communication record between any two members, and the analysis device may Descriptive information is mined to determine the roles of individual members in a given community.
  • the implementation environment may include an analysis device 101 and a communication network 102, where the communication network 102 includes a plurality of member devices 103 through which a plurality of member devices 103 communicate.
  • the analyzing device 101 can monitor the communication behavior of each member device 103, thereby generating the description information.
  • the member device 103 can include a plurality of types of devices, such as a mobile phone, a notebook, and a personal computer. Each member can log in to the member device 103 based on the user identifier, thereby using the member device 103 and other members. To communicate, the user identifier can be a user account, a phone number, or the like.
  • the embodiment of the present application is applied to the mining of the member roles in the specified community, and the hierarchical analysis of the social groups is realized by mining the member roles of the designated group.
  • the working mode of the group is applied to the mining of the member roles in the specified community, and the hierarchical analysis of the social groups is realized by mining the member roles of the designated group.
  • the target group may be a criminal group such as a pyramid scheme group or a money laundering group.
  • the method provided in the embodiment of the present application may be used to mine a plurality of groups, select a target group from the target group, and implement timely reporting and supervision of the target group.
  • the execution body of the method is an analysis device. Referring to FIG. 2, the method includes:
  • any two members can communicate with each other to generate a communication record.
  • the analysis device obtains the communication records of multiple members, the multiple members can be used as a group. Different member roles in the group. Of course, for members who are tentatively assigned to the same role after mining, the analysis device can also use these members as a new group to continue mining the new group and further refine the member roles.
  • the embodiment of the present application is only an example of a specified group, which may be a group initially determined by the analysis device, or may be a group determined after mining the initially determined group.
  • the specified community includes multiple members, and there is a communication record between any two members.
  • a virtual network is created according to the member communication records of the specified group, a corresponding node is generated for each member.
  • a communication record between two members it means that the two members communicated, and the nodes corresponding to the two members are connected to form an edge.
  • the two members are represented. If there is no communication, there is no need to connect the nodes corresponding to the two members.
  • the created virtual network includes multiple nodes, each node represents one member of the specified community, and the two nodes corresponding to the two members communicated form an edge, and the two members that have not communicated The corresponding two nodes are not connected.
  • the virtual network is formed by connecting a plurality of nodes, and the degree of the node refers to the number of nodes directly connected to the node in the virtual network.
  • the core sub-network is less than the pre-removal in the virtual network. Set the number of nodes of the core and the network remaining after the edge of the node that is less than the preset number of cores.
  • the preset number of cores can be determined according to requirements. For example, a small group with a small number of members will set a smaller preset number of presets, and a large group with a larger number of members will set a larger preset number of cores. .
  • the k-core of a sub-network refers to the sub-network remaining after the nodes with less than k degrees and their connected edges are repeatedly removed, and the number of nodes is verified. Indicates the number of cores of the deepest subnetwork containing the node, that is, the node exists in the subnetwork of the k core, but when it does not exist in the subnetwork of the k+1 core, the number of cores of the node is k.
  • the number of cores can be set to 1, then the virtual network itself is a sub-network of 1 core, and then the number of cores is set to 2, and nodes with less than 2 degrees and edges connected by the nodes are removed from the virtual network. Obtain a sub-network of 2 cores, and so on, until a sub-network of k-cores is obtained.
  • the virtual network (1 core network) may be as shown in FIG. 3.
  • the degree of node 1, node 4, node 10, and node 11 is 1, and the degree of other nodes is greater than 1, node 1, node 4, and node 10 are removed.
  • a 2-core subnetwork is obtained.
  • the degree of node 9 becomes 2, and the degree of node 5 and node 6 is also 2, so after removing node 5, node 6, and node 9 and the edges of these nodes, a 3-core subnetwork is obtained.
  • the adjacency matrix uses nodes in the core sub-network as rows and columns.
  • the order of row nodes is the same as the order of column nodes.
  • the order of nodes can be determined randomly, and only the nodes on the rows and columns are arranged.
  • the order is the same, and accordingly, the members are arranged in the same order.
  • the elements at each position in the adjacency matrix are determined according to the connection relationship between the row node corresponding to each location and the column node in the core subnetwork, that is, according to the communication between the corresponding row member and the column member.
  • the row node corresponding to the location when the row node corresponding to the location is connected to the column node, it indicates that the row member corresponding to the location communicates with the column member, and the location is The element is set to the third element; when the row node corresponding to the position is not connected to the column node, it indicates that the row member corresponding to the position has not communicated with the column member, and The element at this position is set to the fourth element.
  • the third element and the fourth element are different elements, and the third element and the fourth element can express differences in relationships between the members.
  • the third element may be larger than the fourth element to indicate that the correlation between the connected nodes is large, and the correlation between the unconnected nodes is small.
  • the third element is 1, and the fourth element is 0.
  • the core sub-network includes four nodes, and the connection relationship is as shown in FIG.
  • the adjacency matrix created according to the core subnetwork can be as follows:
  • the weight of each edge in the virtual network is set to the number of communications or communication parameters between two members corresponding to the two nodes connected by the edge, and then, when the adjacency matrix is created, Set the element at each position in the adjacency matrix to the weight of the edge formed by the corresponding row node and column node connection, that is, set the element at each position to the number of communication between the corresponding row member and column member. Or communication parameters.
  • the number of communications between the two members can represent the correlation between the two, so the elements in each location can be set to correspond between the row members and the column members.
  • Communication messages sent between any two members usually carry communication parameters, which can also represent the correlation between the two, so the elements at each location can be set to the corresponding row members and column members.
  • the communication parameter between the two if a plurality of communication messages are sent between the row member and the column member, the communication parameter between the two may be the sum of the communication parameters carried by the plurality of communication messages.
  • the types of communication parameters carried are different for different types of communication messages. For example, when a payment message is sent between two members, the communication parameter carried is the payment amount, and when the voice message is sent between the two members, the communication parameter carried is the voice duration.
  • the created adjacency matrix can be as 0 M, 14
  • i and j are positive integers less than 5, i, ⁇ , ⁇ 0;
  • the correlation coefficient between each row or each column needs to be repeatedly calculated until the specified matrix is obtained during convergence, but the element on the diagonal of the adjacency matrix is 0, which may result in failure to calculate.
  • the correlation coefficient in turn leads to the inability to obtain the specified matrix.
  • the edge connected to the local node is added to each node in the core sub-network, that is, the self-edge, the simulated node and the node itself perform
  • the communication scenario ensures that when the adjacency matrix is created according to the core sub-network, the elements on the diagonal in the adjacency matrix are greater than 0, thereby ensuring that the standard deviation of each row or column is not zero, thereby ensuring the correlation of sparse adjacency matrix. The validity of the calculation.
  • the default weight of each self-edge is 1, and the obtained adjacency matrix is as follows:
  • the following formula can be used to calculate the correlation coefficient: , , Cov (x, y) .
  • Cov ( X , y) is the covariance of x and ⁇ , . ( ) is the variance, . ( ⁇ is the variance.
  • the first dimension is a row or a column, the first element may be 1, and the second element may be 0.
  • the first dimension is a row
  • the correlation coefficient between any two nodes in the row is calculated, and the obtained correlation coefficients are combined into a correlation coefficient matrix, and then the same calculation method is used.
  • the next correlation coefficient matrix of the correlation coefficient matrix, and so on, until the convergence, the specified matrix including only the first element and the second element is obtained.
  • the specified matrix is obtained in the same manner as described above.
  • the embodiment of the present application repeatedly calculates the correlation coefficient matrix, which is a top-down iterative correlation convergence algorithm, which is based on the following facts: If the correlation coefficients of each row or column in a matrix are repeatedly calculated, finally What will be produced is a correlation coefficient matrix consisting only of the first element and the second element, which realizes the purpose of partitioning each member of the network.
  • the specified matrix includes only the first element and the second element, and the first element is greater than the second element, and the first element indicates that the corresponding row member has stronger correlation with the column member, and the two should belong to the same role, and the second The element indicates that the corresponding row member has weaker relevance to the column member, and the two should belong to different roles. Therefore, according to the first element and the second element in the specified matrix, members in the specified group can be grouped, correlation. A stronger member belongs to the same group, and a less relevant member belongs to a different group, thereby dividing the designated group into a first group and a second group.
  • the communication behavior feature includes a first communication number and a second communication number, the first communication number is a communication number with a member of the same community, and the second communication number is a member of another group The number of communications between.
  • the step 206 can include one or more of the following three situations:
  • the number of communications between each member of the group and members of the same group can be calculated, and the sum of the number of communications between the members of the group and the members of the same group is taken as the first The number of communications, or the average of the number of communications between the members of the same group and the members of the same community as the first number of communications.
  • the first preset number of times and the second preset number of times may be determined according to the manner of calculating the first communication number and the second communication number, wherein the first preset number of times and the second preset number of times may be determined experimentally.
  • the first communication number is greater than the first preset number, indicating that the members of the community have more communication with the members of the same group, and the second communication times are greater than the second preset number, indicating communication between members of the group and members of other groups. More.
  • core members communicate with members inside the group and members outside the group. Therefore, when it is determined that there is more communication between a group and members of the same group, and there are more communication with members of other groups, it can be determined.
  • Members of the group are core members.
  • the contact person belongs to the assisted position in the entire designated group, needs to assist the core member to work, communicates frequently with the assisted members, and there is less communication between the contacts. Therefore, the number of communications with members of the group is less than the number of communications with members of the group. Therefore, when it is determined that the number of communications between members of a certain group and members of the group is less than the number of communications with members outside the group, and there is more communication with members outside the group, it can be determined that the members in the group are contacts. .
  • the middleman is located in the middle of the specified group. It needs to communicate with the members of the upper layer as well as the members of the lower layer. The middlemen have less contact with each other, so there is less communication between the members of the group. More communication with external members of the group. Therefore, when it is determined that the number of communications between a group and a member of a group is small, and the number of communications with members outside the group is large, it can be determined that the members in the group are intermediate.
  • the communication behavior feature includes a third communication number and a fourth communication number
  • the third communication number is the number of times the communication message is sent to the members of the other community
  • the fourth communication number is the member of the other group. The number of times a communication message was sent.
  • the step 206 can include one or more of the following three situations:
  • the communication between any two members is directional, and one member needs to send a communication message to another member, and the communication direction can reflect the hierarchical relationship between the two members to a certain extent.
  • the number of communication times for each member of the group to send a communication message to the members of other groups, and the sum of the number of communication times for all members of the group to send communication messages to members of other groups can be calculated.
  • the third communication number or the group The average of the number of communication times that all members in the body send communication messages to members of other communities as the third communication number.
  • the preset difference value can be determined according to the manner of calculating the third communication number and the fourth communication number, and the preset difference value can be determined experimentally.
  • the underlying members In the group, the underlying members usually report to the core members, so the underlying members send more communication messages, and the core members receive more communication messages. Therefore, when it is determined that members of a certain group send communication to other groups. There are many messages, and the communication messages received by other groups are small. When the difference between the two is large, it can be determined that the members in the group are the underlying members. When it is determined that members of a certain group send less communication messages to other groups, and more communication messages are received by other groups, when the number of the two groups differs greatly, it can be determined that the members in the group are core members.
  • the middleman is located in the middle of the specified community. It only needs to send communication messages to the core members, and also receives communication messages sent by the underlying members, showing that the number of times the communication messages are received is not much different from the number of times the communication messages are sent. Therefore, when it is determined that the number of communications by a member of a certain group to send a communication message to a member of another group is not much different from the number of communications of a communication message sent by a member of another group, it can be determined that the member within the group is a middle person.
  • a directed virtual network may be created, and any two connected nodes in the directed virtual network form an edge, and The direction of this edge matches the direction of communication, sent by The node of the communication message points to the node that receives the communication message.
  • the elements at each position in the adjacency matrix can be divided according to the communication direction, for example, when the row node corresponding to the position points to the column node, The element is set to the third element.
  • the element at the position is set as the fourth element, and then the mining may be performed according to the adjacency matrix, thereby The influence of the communication direction is considered in determining the first group and the second group.
  • the operation flow of the embodiment of the present application may be as shown in FIG. 4C.
  • the core sub-network may be extracted after the core sub-network is extracted. Add weights from edges and edges, and implement member role analysis based on the core subnet.
  • the related criminal network exploration technology focuses on the division of sub-networks in the network, so as to extract core members, but this will cause other important nodes in the sub-network to be ignored, and after extracting to the core members, these core members cannot be Further refinement.
  • the method provided by the embodiment of the present application obtains a correlation matrix by repeatedly calculating a correlation coefficient between each node by acquiring a core sub-network in a virtual network of a specified community, and then obtaining a specified matrix according to the first element and the first matrix in the specified matrix.
  • the second element divides the specified community into a first group and a second group to determine the member roles of each group.
  • the embodiments of the present application can excavate members of at least two roles, expand the scope of excavation, and improve the coverage of the excavation results.
  • the divided groups can be further mined, the role of the member is further refined, the depth of the mining is improved, and the information in the related technology is lost and the data cannot be performed.
  • the problem of role analysis especially the mining and analytic analysis of the target groups in anti-money laundering or anti-MLM projects, and finally achieve tiered review and risk reporting of these groups, greatly improving the efficiency of the review and increasing the risk reporting. Accuracy.
  • FIG. 5 is a flowchart of a method for determining a member role according to an embodiment of the present application. Referring to FIG. 5, the method includes:
  • 501 Create a virtual network according to a member communication record of the specified community.
  • the virtual network includes multiple nodes, each node represents one member of the specified community, and two nodes corresponding to the two members communicated form an edge.
  • 502. Obtain a core sub-network in a virtual network, where the core sub-network is a network that has a degree of removal less than a preset core number and a node that is less than a node of a preset number of cores.
  • the adjacency matrix uses the nodes in the core sub-network as rows and columns, and the row order of the row nodes and the column nodes are arranged in the same order, and the elements in each position in the adjacency matrix are corresponding according to The connection relationship between the row node and the column node is determined.
  • the specified matrix divide the specified community into a first group and a second group, where the row member and the column member corresponding to the first element in the specified matrix are located in the same group in the first group and the second group, and the second element The corresponding row member and column member are located in different groups in the first group and the second group.
  • the method provided by the embodiment of the present application obtains a correlation matrix by repeatedly calculating a correlation coefficient between each node by acquiring a core sub-network in a virtual network of a specified community, and then obtaining a specified matrix according to the first element and the first matrix in the specified matrix.
  • the second element divides the specified community into a first group and a second group to determine the member roles of each group.
  • the embodiments of the present application can excavate members of at least two roles, expand the scope of excavation, and improve the coverage of the excavation results.
  • the divided groups can be further excavated to further refine the member roles and improve the depth of the excavation.
  • the method before the adjacency matrix is created according to the core sub-network, the method further includes:
  • the core sub-network add the edge connected to the local node for each node to ensure that the elements on the diagonal in the adjacency matrix are greater than 0 when the adjacency matrix is created according to the core sub-network.
  • the adjacency matrix is created according to the core sub-network, including: for each position in the adjacency matrix, when the row node corresponding to the location is connected to the column node When the element on the position is set to the third element;
  • the element at the position is set to the fourth element.
  • the method before the adjacency matrix is created according to the core sub-network, the method further includes:
  • the elements at each position in the adjacency matrix are set to the weights of the edges formed by the corresponding row node and column node connections.
  • the communication behavior feature includes a first communication number and a second communication number, the first communication number is the number of communication between the members of the same group, and the second communication number is a member of the other group. Number of communications between;
  • the communication behavior feature includes a third communication number and a fourth communication number
  • the third communication number is the number of times the communication message is sent to the members of the other community
  • the fourth communication number is the member of the other group. The number of times the communication message was sent
  • Determining the member roles of the first group and the second group according to the communication behavior characteristics of the members of the first group and the second group including: When the third communication number of any one of the first group and the second group is greater than the fourth communication number, and the difference between the third communication number and the fourth communication number is greater than a preset difference, determining any group Members are the underlying members;
  • the difference d between the third communication number and the fourth communication number of any one of the first group and the second group is a preset difference, it is determined that the members in any group are intermediate persons.
  • FIG. 6 is a schematic structural diagram of a member role determining apparatus according to an embodiment of the present application.
  • the device includes:
  • the virtual network creation module 601 is configured to create a virtual network according to the member communication record of the specified community, where the virtual network includes multiple nodes, each node represents one member of the specified community, and two nodes corresponding to any two members communicated Connections form an edge;
  • the sub-network obtaining module 602 is configured to obtain a core sub-network in the virtual network, where the core sub-network is a network that has a degree of removal less than a preset number of cores and a network that is less than a node of a preset number of cores;
  • the matrix creation module 603 is configured to create an adjacency matrix according to the core sub-network.
  • the adjacency matrix uses nodes in the core sub-network as rows and columns, and the row-array arrangement order and the column nodes are arranged in the same order, and each position in the adjacency matrix The elements are determined according to the connection relationship between the corresponding row node and the column node;
  • a calculation module 604 configured to: start from the adjacency matrix, repeatedly calculate a correlation coefficient between each node in the first dimension, until a specified matrix including only the first element and the second element is obtained, where the first dimension is a row or a column, One element is larger than the second element;
  • the dividing module 605 is configured to divide the specified group into the first group and the second group according to the specified matrix, where the row member and the column member corresponding to the first element in the specified matrix are located in the same group in the first group and the second group, And the row member and the column member corresponding to the second element are located in different groups in the first group and the second group;
  • the determining module 606 is configured to determine the member roles of the first group and the second group according to the communication behavior characteristics of the members of the first group and the second group.
  • the device further includes:
  • the matrix creation module 603 is configured to:
  • the element at the position is set as the third element
  • the element at the position is set to the fourth element.
  • the device further includes:
  • a weight setting module configured to set a weight of each edge in the core sub-network to a communication number or a communication parameter between two members corresponding to the two nodes connected;
  • the matrix creation module 603 is configured to set an element at each position in the adjacency matrix to a weight of an edge formed by the connection of the corresponding row node and the column node.
  • the communication behavior feature includes a first communication number and a second communication number, the first communication number is the number of communication between the members of the same group, and the second communication number is a member of the other group.
  • the number of times of communication is determined by the determining module 606, configured to: when the first communication number of any one of the first group and the second group is greater than the first preset number of times, and the second communication number is greater than the second preset number of times, determining Members of any group are core members;
  • the communication behavior feature includes a third communication number and a fourth communication number
  • the third communication number is the number of times the communication message is sent to the members of the other community
  • the fourth communication number is the member of the other group. Number of communication messages sent; determination module 606, used to:
  • the difference d between the third communication number and the fourth communication number of any one of the first group and the second group is a preset difference, it is determined that the members in any group are intermediate persons.
  • the member role determining apparatus when determining the member role, the member role determining apparatus provided in the foregoing embodiment is only illustrated by the division of each functional module. In actual applications, the function allocation may be completed by different functional modules as needed. , the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.
  • the member role determining apparatus and the member role determining method embodiment provided in the foregoing embodiments are in the same concept, and the specific implementation process is described in detail in the method embodiment, and details are not described herein again.
  • FIG. 7 is a schematic structural diagram of a terminal according to an embodiment of the present application.
  • the terminal can use the function. Specifically:
  • the terminal 700 may include an RF (Radio Frequency) circuit 110, a memory 120 including one or more computer readable storage media, an input unit 130, a display unit 140, a sensor 150, an audio circuit 160, a transmission module 170, including One or more processing core processor 180, and power supply 190 and the like.
  • RF Radio Frequency
  • the RF circuit 110 can be used for transmitting and receiving information or during a call, and receiving and transmitting the signal. Specifically, after receiving the downlink information of the base station, the downlink information is processed by one or more processors 180. In addition, the data related to the uplink is sent to the base station. .
  • the memory 120 can be used to store software programs and modules, such as the software programs corresponding to the terminals shown in the above exemplary embodiments. And a module, the processor 180 executes various functional applications and data processing by running software programs and modules stored in the memory 120, such as implementing member role determination and the like.
  • the memory 120 may also include a memory controller to provide access to the memory 120 by the processor 180 and the input unit 130.
  • Input unit 130 can be used to receive input numeric or character information, as well as to generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.
  • input unit 130 can include touch-sensitive surface 131 as well as other input terminals 132.
  • Display unit 140 can be used to display information entered by the user or information provided to the user and various graphical user interfaces of terminal 700, which can be constructed from graphics, text, icons, video, and any combination thereof.
  • Terminal 700 can also include at least one type of sensor 150, such as a light sensor, motion sensor, and other sensors.
  • the audio circuit 160, the speaker 161, and the microphone 162 provide an audio interface between the user and the terminal 700.
  • the terminal 700 can help the user to send and receive emails, browse web pages, access streaming media, etc. through the transmission module 170, which provides users with wireless or wired broadband Internet access.
  • the processor 180 is the control center of the terminal 700, which links various portions of the entire handset using various interfaces and lines, by running or executing software programs and/or modules stored in the memory 120, and recalling data stored in the memory 120, The various functions and processing data of the terminal 700 are performed to perform overall monitoring of the mobile phone.
  • Processor 180 can include one or more processing cores.
  • Terminal 700 also includes a power source 190 (such as a battery) that powers the various components.
  • a power source 190 such as a battery
  • the terminal 700 further includes a memory and at least one instruction, at least one program, a code set or a set of instructions, wherein at least one instruction, at least one program, code set or instruction set is stored in the memory and configured to be processed by one or more Loaded and executed to create a virtual network based on the member communication records of the specified community;
  • the member roles of the first group and the second group are determined according to the communication behavior characteristics of the members of the first group and the second group.
  • FIG. 8 is a schematic structural diagram of a server according to an embodiment of the present application.
  • the server 800 may have a large difference due to different configurations or performances, and may include one or more central processing units (CPUs) 822 ( For example, one or more processors) and memory 832, one or more storage media 830 that store application 842 or data 844 (eg, one or one storage device in Shanghai).
  • the memory 832 and the storage medium 830 may be short-term storage or persistent storage.
  • Programs stored on storage medium 830 may include one or more modules (not shown), and each module may include a series of instruction operations on the server.
  • Server 800 may also include one or more power supplies 826, one or more wired or wireless network interfaces 850, one or more input and output interfaces 858, one or more keyboards 856, and/or one or more operating systems 841 .
  • the central processing unit 822 can be configured to communicate with the storage medium 830, load a series of instruction operations in the storage medium 830, and execute the above-mentioned member role determination method on the server 800, including:
  • the central processing unit 822 can load a series of instruction operations in the storage medium 830, and implement an operation of adding an edge connected to the local node to each node on the server 800, or setting each of the adjacency matrix. The operation of the elements in the position, etc., will not be described here.
  • the embodiment of the present application further provides a member role determining device, where the member role determining device includes a processor and a memory, where the memory stores at least one instruction, at least one program, code set or instruction set, instruction, program, code set or The instruction set is loaded and executed by the processor.
  • the embodiment of the present application further provides a computer readable storage medium, where the computer readable storage medium stores at least one instruction, at least one program, a code set or a set of instructions, the instruction, The program, the set of codes, or the set of instructions is loaded and executed by the processor to implement the operations performed in the member role determining method of the above-described embodiments.
  • a person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium.
  • the storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本申请实施例公开了一种成员角色确定方法,包括:根据指定团体的成员通信记录创建虚拟网络;获取虚拟网络中的核心子网络;根据核心子网络创建邻接矩阵;从邻接矩阵开始,重复计算第一维度上各个节点之间的相关系数,直至获取到仅包括第一元素和第二元素的指定矩阵;根据指定矩阵将指定团体划分为第一团体和第二团体,确定第一团体和第二团体的成员角色。

Description

成员角色确定方法、 装置及存储介质 本申请要求于 2017年 8月 30日提交中国专利局、 申请号为 201710761789.6, 发明名称为"成员角色确定方法、 装置及存储介质"的 中国专利申请的优先权, 其全部内容通过引用结合在本申请中。 技术领域 本申请实施例涉及数据挖掘领域, 特别涉及一种成员角色确定方 法、 装置及存储介质。 发明背景 现实生活中存在着各种各样的社会团体, 诸如家庭团体、 交易团体 等, 这些社会团体中包括多个成员, 每个成员扮演着不同的角色, 如何 确定成员的角色已成为与社会生活联系最紧密的问题。
相关的挖掘方法中, 通常先根据指定团体的成员通信记录创建虚拟 网络,该虚拟网络包括多个节点,每个节点代表指定团体中的一个成员, 通信过的任两个成员所对应的两个节点连接构成一条边。 之后, 釆用层 次聚类算法对该虚拟网络中的成员进行聚类, 得到该虚拟网络中包含的 多个子网络, 釆用 Spline (样条插值)算法计算每个子网络中每个节点 的中心度, 中心度可以代表成员控制的以自己为中心的子网络的紧密程 度, 并将中心度大于预设次数的节点对应的成员确定为指定团体的核心 成员。
除了核心成员之外, 指定团体中还存在着联络人、 中间人等其他角 色的成员, 釆用上述挖掘方法仅能挖掘出核心成员这一种角色的成员, 而不能挖掘出其他角色的成员。 因此上述挖掘方法存在局限性, 挖掘结 果不够全面。 发明内容 为了解决相关技术的问题, 本申请实施例提供了一种成员角色确定 方法、 装置及存储介质。 所述技术方案如下:
第一方面, 提供了一种成员角色确定方法, 应用于分析设备, 所述 方法包括:
根据指定团体的成员通信记录创建虚拟网络, 所述虚拟网络包括多 个节点, 每个节点代表所述指定团体中的一个成员, 通信过的任两个成 员所对应的两个节点连接构成一条边;
获取所述虚拟网络中的核心子网络, 所述核心子网络为所述虚拟网 络中去除度小于预设核数的节点以及所述度小于预设核数的节点连接 的边之后剩余的网络;
居所述核心子网络创建邻接矩阵, 所述邻接矩阵以所述核心子网 络中的节点作为行和列, 行节点的排列顺序和列节点的排列顺序相同, 且所述邻接矩阵中每个位置上的元素根据对应的行节点与列节点之间 的连接关系确定;
从所述邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系 数, 直至获取到仅包括第一元素和第二元素的指定矩阵, 所述第一维度 为行或者列, 所述第一元素大于所述第二元素;
根据所述指定矩阵, 将所述指定团体划分为第一团体和第二团体, 其中所述指定矩阵中所述第一元素对应的行成员和列成员位于所述第 一团体和所述第二团体中的同一团体, 而所述第二元素对应的行成员和 列成员位于所述第一团体和所述第二团体中的不同团体;
根据所述第一团体和所述第二团体的成员的通信行为特征, 确定所 述第一团体和所述第二团体的成员角色。
第二方面, 提供了一种成员角色确定装置, 应用于分析设备, 所述 装置包括处理器和存储器, 所述存储器中存储可被所述处理器执行的指 令, 当执行所述指令时, 所述处理器用于:
根据指定团体的成员通信记录创建虚拟网络, 所述虚拟网络包括多 个节点, 每个节点代表所述指定团体中的一个成员, 通信过的任两个成 员所对应的两个节点连接构成一条边;
获取所述虚拟网络中的核心子网络, 所述核心子网络为所述虚拟网 络中去除度小于预设核数的节点以及所述度小于预设核数的节点连接 的边之后剩余的网络;
根据所述核心子网络创建邻接矩阵, 所述邻接矩阵以所述核心子网 络中的节点作为行和列, 行节点的排列顺序和列节点的排列顺序相同, 且所述邻接矩阵中每个位置上的元素根据对应的行节点与列节点之间 的连接关系确定;
从所述邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系 数, 直至获取到仅包括第一元素和第二元素的指定矩阵, 所述第一维度 为行或者列, 所述第一元素大于所述第二元素;
根据所述指定矩阵, 将所述指定团体划分为第一团体和第二团体, 其中所述指定矩阵中所述第一元素对应的行成员和列成员位于所述第 一团体和所述第二团体中的同一团体, 而所述第二元素对应的行成员和 列成员位于所述第一团体和所述第二团体中的不同团体;
根据所述第一团体和所述第二团体的成员的通信行为特征, 确定所 述第一团体和所述第二团体的成员角色。
第三方面, 提供了一种计算机可读存储介质, 所述计算机可读存储 介质中存储有至少一条指令、 至少一段程序、 代码集或指令集, 所述指 令、 所述程序、 所述代码集或所述指令集由处理器加载并执行以实现如 第一方面所述的成员角色确定方法中所执行的操作。 附图简要说明 为了更清楚地说明本申请实施例中的技术方案, 下面将对实施例描 述中所需要使用的附图作简单地介绍, 显而易见地, 下面描述中的附图 仅仅是本申请的一些实施例, 对于本领域普通技术人员来讲, 在不付出 创造性劳动的前提下, 还可以根据这些附图获得其他的附图。
图 1是本申请实施例提供的一种实施环境的示意图;
图 2是本申请实施例提供的一种成员角色确定方法的流程图; 图 3是本申请实施例提供的一种核心子网络的示意图;
图 4 A是本申请实施例提供的一种核心子网络的示意图;
图 4B是本申请实施例提供的一种添加自边的示意图;
图 4C是本申请实施例提供的一种操作流程示意图;
图 5是本申请实施例提供的一种成员角色确定方法的流程图; 图 6是本申请实施例提供的一种成员角色确定装置的结构示意图; 图 7是本申请实施例提供的一种终端的结构示意图;
图 8是本申请实施例提供的一种服务器的结构示意图。 实施方式 下面将结合本申请实施例中的附图, 对本申请实施例中的技术方案 进行清楚、 完整地描述, 显然, 所描述的实施例是本申请一部分实施例, 而不是全部的实施例。 基于本申请中的实施例, 本领域普通技术人员在 没有做出创造性劳动前提下所获得的所有其他实施例, 都属于本申请保 护的范围。
本申请实施例的实施环境可以包括分析设备, 该分析设备获取指定 团体的描述信息, 该描述信息中包括指定团体中的多个成员以及任两个 成员之间的通信记录, 分析设备可以根据该描述信息进行挖掘, 从而确 定指定团体中各个成员的角色。
在一种可能实现方式中, 参见图 1, 该实施环境可以包括分析设备 101和通信网络 102, 该通信网络 102中包括多个成员设备 103, 多个成 员设备 103通过该通信网络 102进行通信的过程中, 分析设备 101可以 监控各个成员设备 103的通信行为, 从而生成该描述信息。
其中, 分析设备可以为计算机或者服务器等, 成员设备 103可以包 括手机、 笔记本、 个人电脑等多种类型的设备, 每个成员可以基于用户 标识登录成员设备 103,从而通过该成员设备 103与其他成员进行通信, 该用户标识可以为用户账号、 电话号码等。
本申请实施例应用于对指定团体中的成员角色进行挖掘的场景中, 通过挖掘指定团体的成员角色, 实现了对社会团体的层次分析, 了解社 会团体的工作模式。
并且, 通过挖掘指定团体的成员角色, 可以判定该指定团体是否为 目标团体。 其中, 该目标团体可以为传销团体、 洗钱团体等犯罪团体, 釆用本申请实施例提供的方法, 可以对多种团体进行挖掘, 从中筛选出 目标团体, 实现对目标团体的及时上报和监管。
图 2是本申请实施例提供的一种成员角色确定方法的流程图, 该方 法的执行主体为分析设备, 参见图 2, 该方法包括:
201、 获取指定团体的成员通信记录, 根据获取的成员通信记录创 建虚拟网络。
实际应用中, 任两个成员之间可以进行通信, 生成通信记录, 当分 析设备获取到多个成员的通信记录时, 可以先将该多个成员作为一个团 体, 对该团体进行挖掘可以确定该团体中不同的成员角色。 当然, 对于 挖掘后暂定为属于同一角色的成员, 分析设备也可以将这些成员作为一 个新的团体, 继续对该新的团体进行挖掘, 实现对成员角色的进一步细 化。
本申请实施例仅以指定团体为例, 该指定团体可以为分析设备初始 确定的团体, 也可以为对初始确定的团体进行挖掘后所确定的团体。
针对该指定团体来说, 该指定团体中包括多个成员, 任两个成员之 间存在通信记录, 根据该指定团体的成员通信记录创建虚拟网络时, 为 每个成员生成对应的节点, 当任两个成员之间存在通信记录时, 表示这 两个成员通信过, 则将这两个成员对应的节点连接构成一条边, 当任两 个成员之间不存在通信记录时, 表示这两个成员没有通信过, 则无需将 这两个成员对应的节点连接。
相应地, 创建的虚拟网络中包括多个节点, 每个节点代表指定团体 中的一个成员, 通信过的任两个成员所对应的两个节点连接构成一条 边, 而未通信过的两个成员所对应的两个节点不连接。
202、 获取虚拟网络中的核心子网络。
其中, 虚拟网络由多个节点连接而成, 节点的度是指虚拟网络中与 该节点直接连接的节点的个数。 核心子网络为虚拟网络中去除度小于预 设核数的节点以及度小于预设核数的节点连接的边之后剩余的网络。 该 预设核数可以根据需求确定, 例如通常情况下, 针对成员数量较小的小 型团体会设置较小的预设核数, 针对成员数量较大的大型团体会设置较 大的预设核数。
假设预设核数为 k, k为正整数, k核子网络的定义为: 一个子网络 的 k核是指反复去掉度小于 k的节点及其连接的边后剩余的子网络, 节 点的核数表示包含该节点的最深的子网络的核数, 即节点存在于 k核的 子网络, 但不存在于 k+1核的子网络时, 节点的核数为 k。
实际应用时, 可以先将核数设置为 1, 则虚拟网络本身即为 1核的 子网络, 之后将核数设置为 2, 从虚拟网络中去除度小于 2的节点以及 这些节点连接的边, 得到 2核的子网络, 以此类推, 直至得到 k核的子 网络。
例如, 该虚拟网络( 1核子网络)可以如图 3所示, 节点 1、 节点 4、 节点 10和节点 11的度为 1, 其他节点的度大于 1, 则去除节点 1、 节点 4、 节点 10和节点 11 以及这些节点连接的边后, 得到 2核子网络。 此 时, 节点 9的度变为 2, 节点 5和节点 6的度也为 2, 因此去除节点 5、 节点 6和节点 9以及这些节点连接的边后, 得到 3核子网络。
203、 根据核心子网络创建邻接矩阵。
其中, 邻接矩阵以核心子网络中的节点作为行和列, 行节点的排列 顺序和列节点的排列顺序相同, 实际应用时节点的排列顺序可以随机确 定, 只需保证行和列上的节点排列顺序相同即可, 相应地, 所代表的成 员排列顺序也相同。
邻接矩阵中每个位置上的元素根据每个位置对应的行节点与列节 点在核心子网络中的连接关系确定, 也即是根据对应的行成员与列成员 之间的通信 i己录确定。
在第一种可能实现方式中, 对于邻接矩阵中的每个位置, 当该位置 对应的行节点与列节点连接时, 表示该位置对应的行成员与列成员通信 过, 此时将该位置上的元素设置为第三元素; 当该位置对应的行节点与 列节点未连接时, 表示该位置对应的行成员与列成员未通信过, 此时将 该位置上的元素设置为第四元素。
其中, 该第三元素和该第四元素为不同的元素, 通过第三元素和第 四元素可以表现出成员之间关系的不同。 该第三元素可以大于该第四元 素, 以表示连接的节点之间的相关性较大, 未连接的节点之间的相关性 较小。 例如该第三元素为 1, 该第四元素为 0。
例如, 参见图 4A, 该核心子网络包括四个节点, 且连接关系如图
4A所示。 根据该核心子网络所创建的邻接矩阵可以如下:
M = ;
Figure imgf000009_0001
其中, i和 j为小于 5的正整数, i = j时, Mij = 0 ; i≠ j时, ¾^表 示节点 i与节点 j之间是否连接, ¾^ = 1表示节点 i与节点 j之间连接, ¾^ = 0表示节点 i与节点 j之间不连接。
在第二种可能实现方式中, 将虚拟网络中的每条边的权重设置为这 条边连接的两个节点所对应的两个成员之间的通信次数或通信参数, 那 么在创建邻接矩阵时将邻接矩阵中每个位置上的元素设置为对应的行 节点与列节点连接构成的边的权重, 也即是将每个位置上的元素设置为 对应的行成员与列成员之间的通信次数或通信参数。
任两个成员之间进行通信时, 这两个成员之间的通信次数可以代表 两者之间的相关性, 因此可以将每个位置上的元素设置为对应的行成员 与列成员之间的通信次数。 任两个成员之间所发送的通信消息中通常会 携带通信参数, 该通信参数也可以代表两者之间的相关性, 因此可以将 每个位置上的元素设置为对应的行成员与列成员之间的通信参数, 如果 行成员与列成员之间发送了多条通信消息时两者之间的通信参数可以 为多条通信消息所携带的通信参数之和。
其中,对于不同类型的通信消息,所携带的通信参数的类型也不同。 例如两个成员之间发送支付消息时, 所携带的通信参数为支付金额, 两 个成员之间发送语音消息时, 所携带的通信参数为语音时长。
例如, 根据如图 4A所示的核心子网络, 所创建的邻接矩阵可以如 0 M, 14
M 0 24
M
M 0 M 34
M 42 M 43 0
其中, i和 j为小于 5的正整数, i 时, Μ,· 0;
节点 i与节点 j之间的通信次数。
本申请实施例中, 考虑到创建邻接矩阵之后需要重复计算每行或者 每列之间的相关系数, 直至收敛时得到指定矩阵, 然而邻接矩阵的对角 线上的元素为 0可能会导致无法计算相关系数进而导致无法获取到指定 矩阵。 为了避免这一问题, 在一种可能实现方式中, 在获取到核心子网 络之后,在核心子网络中为每个节点添加与本端节点连接的边,即自边, 模拟节点与节点本身进行通信的场景, 这样能够保证根据核心子网络创 建邻接矩阵时, 邻接矩阵中对角线上的元素大于 0, 进而保证每行或者 每列的标准差不为 0, 从而保证了稀疏邻接矩阵相关性计算的有效性。
例如, 在图 4A所示的核心子网络中, 为每个节点添加与自身连接 的边, 得到如图 4B所示的核心子网络。 j
才艮据如图 4B所示的核心子网络, 釆用上述第一种可能实现方式创
Ml 建邻接矩阵时, 得到的邻接矩阵如下: 为
1 0 1
0 1 1
M =
1 1 1
0 1 1
才艮据如图 4B所示的核心子网络, 釆用上述第二种可能实现方式创 建邻接矩阵时, 默认每条自边的权重为 1, 得到的邻接矩阵如下:
Figure imgf000010_0001
204、 从邻接矩阵开始, 重复计算第一维度上各个节点之间的相关 系数, 直至获取到仅包括第一元素和第二元素的指定矩阵。
其中, 可以应用以下公式计算相关系数: , 、 Cov (x, y) .
° (x) · 0 (y)
Cov (X, y)为 x与 Υ的协方差, 。( )为 的方差, 。(^为 的方差。 其中, 第一维度为行或者列, 该第一元素可以为 1, 该第二元素可 以为 0。
当第一维度为行时, 对于邻接矩阵中的任一行, 计算这一行中任两 个节点之间的相关系数, 将得到的相关系数组成一个相关系数矩阵, 之 后继续釆用同样的计算方式得到该相关系数矩阵的下一个相关系数矩 阵, 以此类推, 直至收敛时获取到仅包括第一元素和第二元素的指定矩 阵。 当第一维度为列时, 釆用上述同样的方式获取指定矩阵。
本申请实施例重复计算相关系数矩阵, 釆用的是一种自上而下的迭 代相关收敛算法, 该算法基于如下事实: 如果对一个矩阵中的各行或者 各列的相关系数进行重复计算, 最终产生的将是一个仅仅由第一元素和 第二元素组成的相关系数矩阵, 实现了对网络中各个成员进行分区的目 的。
205、 根据指定矩阵, 将指定团体划分为第一团体和第二团体, 其 中指定矩阵中第一元素对应的行成员和列成员位于第一团体和第二团 体中的同一团体, 而第二元素对应的行成员和列成员位于第一团体和第 二团体中的不同团体。
指定矩阵中仅包括第一元素和第二元素, 且第一元素大于第二元 素, 第一元素表示对应的行成员与列成员的相关性较强, 两者应当属于 同一种角色, 而第二元素表示对应的行成员与列成员的相关性较弱, 两 者应当属于不同的角色, 因此根据该指定矩阵中的第一元素和第二元 素, 可以对指定团体中的成员进行分组, 相关性较强的成员属于同一个 团体, 而相关性较弱的成员属于不同的团体, 从而将指定团体划分为第 一团体和第二团体。
206、 根据第一团体和第二团体的成员的通信行为特征, 确定第一 团体和第二团体的成员角色。
得到第一团体和第二团体之后, 可以确定这两个团体的成员角色不 同, 至于每个团体的成员具体是什么角色需要根据成员的通信行为特征 确定。
在第一种可能实现方式中, 通信行为特征包括第一通信次数和第二 通信次数, 第一通信次数为与同一团体的成员之间的通信次数, 第二通 信次数为与其他团体的成员之间的通信次数。
那么, 该步骤 206可以包括以下三种情况的一种或多种:
1-1、当第一团体和第二团体中任一团体的第一通信次数大于第一预 设次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员 为核心成员。
对于每个团体来说, 可以计算该团体内每个成员与同一团体的成员 之间的通信次数, 将该团体内所有成员计算出的、 与同一团体的成员之 间的通信次数之和作为第一通信次数, 或者将该团体内所有成员计算出 的、 与同一团体的成员之间的通信次数的平均值作为第一通信次数。
还可以计算该团体内每个成员与其他团体的成员之间的通信次数, 将该团体内所有成员计算出的、 与其他团体的成员之间的通信次数之和 作为第二通信次数, 或者将该团体内所有成员计算出的、 与其他团体的 成员之间的通信次数的平均值作为第二通信次数。
相应地, 可以根据计算第一通信次数和第二通信次数的方式确定第 一预设次数和第二预设次数, 其中, 该第一预设次数和该第二预设次数 可以通过实验确定, 第一通信次数大于第一预设次数, 表示团体的成员 与同一团体的成员之间通信较多, 第二通信次数大于第二预设次数, 表 示团体的成员与其他团体的成员之间的通信较多。
通常核心成员与团体内部的成员以及团体外部的成员通信都比较 频繁, 因此当确定某一团体与同一团体的成员之间通信较多, 与其他团 体的成员之间通信也较多时, 可以确定该团体内的成员是核心成员。
1-2、当第一团体和第二团体中任一团体的第一通信次数小于第二通 信次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员 为联络人。
通常联络人属于整个指定团体中的协助位置, 需要协助核心成员进 行工作, 与所协助的成员之间通信比较频繁, 而联络人之间通信较少, 因此呈现出与团体内部成员之间的通信次数小于与团体外部成员之间 的通信次数的特点。 因此, 当确定某一团体的成员与团体内部成员之间 的通信次数小于与团体外部成员之间的通信次数, 且与团体外部成员之 间通信较多时, 可以确定该团体内的成员是联络人。
1- 3、当第一团体和第二团体中任一团体的第一通信次数小于第一预 设次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员 为中间人。
通常中间人位于指定团体中的中间位置, 既要与上层的成员进行通 信, 也需要与下层的成员进行通信, 而中间人互相之间联系较少, 因此 呈现出团体内部成员之间的通信较少, 与团体外部成员之间的通信较多 的特点。 因此, 当确定某一团体与团体内部成员之间的通信次数较小, 与团体外部成员之间的通信次数较大时, 可以确定该团体内的成员是中 间人。
在第二种可能实现方式中, 通信行为特征包括第三通信次数和第四 通信次数, 第三通信次数为向其他团体的成员发送通信消息的次数, 第 四通信次数为接收其他团体的成员所发送的通信消息的次数。
那么, 该步骤 206可以包括以下三种情况的一种或多种:
2- 1、当第一团体和第二团体中任一团体的第三通信次数大于第四通 信次数, 且第三通信次数与第四通信次数之间的差值大于预设差值时, 确定任一团体的成员为底层成员。
2-2、当第一团体和第二团体中的任一团体的第四通信次数大于第三 通信次数, 且第四通信次数与第三通信次数之间的差值大于预设差值 时, 确定任一团体内的成员为核心成员。
任两个成员之间的通信具有方向性, 需要由一个成员向另一个成员 发送通信消息, 而通信方向在一定程度上可以体现两个成员之间的层级 关系。
对于每个团体来说, 可以计算该团体内每个成员作为发送方, 向其 他团体的成员发送通信消息的通信次数, 将该团体内所有成员向其他团 体的成员发送通信消息的通信次数之和作为第三通信次数, 或者将该团 体内所有成员向其他团体的成员发送通信消息的通信次数的平均值作 为第三通信次数。
还可以计算该团体内每个成员作为接收方, 接收其他团体的成员发 送的通信消息的通信次数, 将该团体内所有成员接收其他团体成员发送 的通信消息的通信次数之和作为第四通信次数, 或者将该团体内所有成 员接收其他团体成员发送的通信消息的通信次数的平均值作为第四通 信次数。
相应地, 可以根据计算第三通信次数和第四通信次数的方式确定预 设差值, 该预设差值可以通过实验确定。
在团体内通常是由底层成员向核心成员汇报, 因此会呈现出底层成 员发送通信消息较多, 而核心成员接收通信消息较多的特点, 因此, 当 确定某一团体的成员向其他团体发送通信消息较多, 而接收其他团体的 通信消息较小, 两者次数相差较大时, 可以确定该团体内的成员为底层 成员。 而当确定某一团体的成员向其他团体发送通信消息较少, 而接收 其他团体的通信消息较多, 两者的次数相差较大时, 可以确定该团体内 的成员为核心成员。
2-3、当第一团体和第二团体中的任一团体的第三通信次数与第四通 信次数之间的差值小于预设差值时, 确定任一团体内的成员为中间人。
通常中间人位于指定团体中的中间位置, 既要向核心成员发送通信 消息, 也需要接收底层成员发送的通信消息, 呈现出接收通信消息的次 数与发送通信消息的次数相差不大的特点。 因此, 当确定某一团体的成 员向其他团体的成员发送通信消息的通信次数与接收其他团体的成员 发送的通信消息的通信次数相差不大时, 可以确定该团体内的成员为中 间人。
当然, 除上述两种可能实现方式之外, 还可以釆用其他类型的通信 行为特征确定第一团体和第二团体的成员角色。
需要说明的是, 为了便于确定通信方向, 在上述步骤 201-204中, 当创建虚拟网络时可以创建有向虚拟网络, 该有向虚拟网络中任两个通 信过的节点连接构成一条边, 且这条边的方向与通信方向匹配, 由发送 通信消息的节点指向接收通信消息的节点, 在创建邻接矩阵时, 邻接矩 阵中每个位置上的元素可以按照通信方向进行划分, 例如当位置对应的 行节点指向列节点时, 将该位置上的元素设置为第三元素, 当该位置对 应的列节点指向行节点, 或者行节点与列节点未连接时, 将该位置上的 元素设置为第四元素, 后续可以根据该邻接矩阵进行挖掘, 从而在确定 第一团体和第二团体时考虑通信方向的影响。
综上所述,本申请实施例的操作流程可以如图 4C所示,参见图 4C, 对于指定团体, 通过对指定团体的虚拟网络进行深度分析, 提取出核心 子网络后, 可以为核心子网络添加自边以及边的权重, 并才艮据该核心子 网络实现成员角色分析。
相关的犯罪网络的探索技术专注于网络中子网络的划分, 从而进行 核心成员的提取, 但是这会导致子网络中的其他重要节点被忽视, 而且 提取到核心成员之后, 并不能对这些核心成员进行进一步地细化。
本申请实施例提供的方法, 通过获取指定团体的虚拟网络中的核心 子网络, 创建邻接矩阵后, 通过重复计算各个节点之间的相关系数得到 指定矩阵, 根据指定矩阵中的第一元素和第二元素, 将指定团体划分为 第一团体和第二团体, 从而确定每个团体的成员角色。 本申请实施例可 以挖掘出至少两种角色的成员, 扩大了挖掘范围, 提升了挖掘结果的覆 盖面。 尤其是, 重复釆用本申请实施例提供的方法, 可以对划分出的团 体继续进行挖掘, 实现对成员角色的进一步细化, 提高了挖掘深度, 解 决了相关技术中子网络信息流失以及无法进行角色分析的问题, 尤其是 可以在反洗钱或反传销等项目中对目标团体的挖掘和层次分析, 最终实 现对这些团体的分层审核和风险上报, 大大提高了审核效率, 增加了风 险上报的准确率。
图 5是本申请实施例提供的一种成员角色确定方法的流程图, 参见 图 5, 该方法包括:
501、 根据指定团体的成员通信记录创建虚拟网络, 虚拟网络包括 多个节点, 每个节点代表指定团体中的一个成员, 通信过的任两个成员 所对应的两个节点连接构成一条边。 502、 获取虚拟网络中的核心子网络, 核心子网络为虚拟网络中去 除度小于预设核数的节点以及度小于预设核数的节点连接的边之后剩 余的网络。
503、 根据核心子网络创建邻接矩阵, 邻接矩阵以核心子网络中的 节点作为行和列, 行节点的排列顺序和列节点的排列顺序相同, 且邻接 矩阵中每个位置上的元素根据对应的行节点与列节点之间的连接关系 确定。
504、 从邻接矩阵开始, 重复计算第一维度上各个节点之间的相关 系数, 直至获取到仅包括第一元素和第二元素的指定矩阵, 第一维度为 行或者列, 第一元素大于第二元素。
505、 根据指定矩阵, 将指定团体划分为第一团体和第二团体, 其 中指定矩阵中第一元素对应的行成员和列成员位于第一团体和第二团 体中的同一团体, 而第二元素对应的行成员和列成员位于第一团体和第 二团体中的不同团体。
506、 根据第一团体和第二团体的成员的通信行为特征, 确定第一 团体和第二团体的成员角色。
本申请实施例提供的方法, 通过获取指定团体的虚拟网络中的核心 子网络, 创建邻接矩阵后, 通过重复计算各个节点之间的相关系数得到 指定矩阵, 根据指定矩阵中的第一元素和第二元素, 将指定团体划分为 第一团体和第二团体, 从而确定每个团体的成员角色。 本申请实施例可 以挖掘出至少两种角色的成员, 扩大了挖掘范围, 提升了挖掘结果的覆 盖面。 尤其是, 重复釆用本申请实施例提供的方法, 可以对划分出的团 体继续进行挖掘, 实现对成员角色的进一步细化, 提高了挖掘深度。
在一种可能实现方式中, 根据核心子网络创建邻接矩阵之前, 方法 还包括:
在核心子网络中, 为每个节点添加与本端节点连接的边, 以保证根 据核心子网络创建邻接矩阵时, 邻接矩阵中对角线上的元素大于 0。
在另一种可能实现方式中, 根据核心子网络创建邻接矩阵, 包括: 对于邻接矩阵中的每个位置, 当位置对应的行节点与列节点连接 时, 将位置上的元素设置为第三元素;
当位置对应的行节点与列节点未连接时, 将位置上的元素设置为第 四元素。
在另一种可能实现方式中, 根据核心子网络创建邻接矩阵之前, 方 法还包括:
将核心子网络中的每条边的权重设置为连接的两个节点所对应的 两个成员之间的通信次数或通信参数;
根据核心子网络创建邻接矩阵, 包括:
将邻接矩阵中每个位置上的元素设置为对应的行节点与列节点连 接构成的边的权重。
在另一种可能实现方式中, 通信行为特征包括第一通信次数和第二 通信次数, 第一通信次数为与同一团体的成员之间的通信次数, 第二通 信次数为与其他团体的成员之间的通信次数;
根据第一团体和第二团体的成员的通信行为特征, 确定第一团体和 第二团体的成员角色, 包括:
当第一团体和第二团体中任一团体的第一通信次数大于第一预设 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 核心成员;
当第一团体和第二团体中任一团体的第一通信次数小于第二通信 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 联络人;
当第一团体和第二团体中任一团体的第一通信次数小于第一预设 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 中间人。
在另一种可能实现方式中, 通信行为特征包括第三通信次数和第四 通信次数, 第三通信次数为向其他团体的成员发送通信消息的次数, 第 四通信次数为接收其他团体的成员所发送的通信消息的次数;
根据第一团体和第二团体的成员的通信行为特征, 确定第一团体和 第二团体的成员角色, 包括: 当第一团体和第二团体中任一团体的第三通信次数大于第四通信 次数, 且第三通信次数与第四通信次数之间的差值大于预设差值时, 确 定任一团体的成员为底层成员;
当第一团体和第二团体中的任一团体的第四通信次数大于第三通 信次数, 且第四通信次数与第三通信次数之间的差值大于预设差值时, 确定任一团体内的成员为核心成员;
当第一团体和第二团体中的任一团体的第三通信次数与第四通信 次数之间的差值 d、于预设差值时, 确定任一团体内的成员为中间人。
图 6是本申请实施例提供的一种成员角色确定装置的结构示意图。 参见图 6, 该装置包括:
虚拟网络创建模块 601, 用于根据指定团体的成员通信记录创建虚 拟网络,虚拟网络包括多个节点,每个节点代表指定团体中的一个成员, 通信过的任两个成员所对应的两个节点连接构成一条边;
子网络获取模块 602, 用于获取虚拟网络中的核心子网络, 核心子 网络为虚拟网络中去除度小于预设核数的节点以及度小于预设核数的 节点连接的边之后剩余的网络;
矩阵创建模块 603, 用于根据核心子网络创建邻接矩阵, 邻接矩阵 以核心子网络中的节点作为行和列, 行节点的排列顺序和列节点的排列 顺序相同, 且邻接矩阵中每个位置上的元素根据对应的行节点与列节点 之间的连接关系确定;
计算模块 604, 用于从邻接矩阵开始, 重复计算第一维度上各个节 点之间的相关系数, 直至获取到仅包括第一元素和第二元素的指定矩 阵, 第一维度为行或者列, 第一元素大于第二元素;
划分模块 605, 用于根据指定矩阵, 将指定团体划分为第一团体和 第二团体, 其中指定矩阵中第一元素对应的行成员和列成员位于第一团 体和第二团体中的同一团体, 而第二元素对应的行成员和列成员位于第 一团体和第二团体中的不同团体;
确定模块 606, 用于根据第一团体和第二团体的成员的通信行为特 征, 确定第一团体和第二团体的成员角色。 在一种可能实现方式中, 装置还包括:
添加模块, 用于在核心子网络中, 为每个节点添加与本端节点连接 的边, 以保证根据核心子网络创建邻接矩阵时, 邻接矩阵中对角线上的 元素大于 0。
在另一种可能实现方式中, 矩阵创建模块 603用于:
对于邻接矩阵中的每个位置, 当位置对应的行节点与列节点连接 时, 将位置上的元素设置为第三元素;
当位置对应的行节点与列节点未连接时, 将位置上的元素设置为第 四元素。
在另一种可能实现方式中, 装置还包括:
权重设置模块, 用于将核心子网络中的每条边的权重设置为连接的 两个节点所对应的两个成员之间的通信次数或通信参数;
矩阵创建模块 603, 用于将邻接矩阵中每个位置上的元素设置为对 应的行节点与列节点连接构成的边的权重。
在另一种可能实现方式中, 通信行为特征包括第一通信次数和第二 通信次数, 第一通信次数为与同一团体的成员之间的通信次数, 第二通 信次数为与其他团体的成员之间的通信次数; 确定模块 606, 用于: 当第一团体和第二团体中任一团体的第一通信次数大于第一预设 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 核心成员;
当第一团体和第二团体中任一团体的第一通信次数小于第二通信 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 联络人;
当第一团体和第二团体中任一团体的第一通信次数小于第一预设 次数, 且第二通信次数大于第二预设次数时, 确定任一团体内的成员为 中间人。
在另一种可能实现方式中, 通信行为特征包括第三通信次数和第四 通信次数, 第三通信次数为向其他团体的成员发送通信消息的次数, 第 四通信次数为接收其他团体的成员所发送的通信消息的次数; 确定模块 606, 用于:
当第一团体和第二团体中任一团体的第三通信次数大于第四通信 次数, 且第三通信次数与第四通信次数之间的差值大于预设差值时, 确 定任一团体的成员为底层成员;
当第一团体和第二团体中的任一团体的第四通信次数大于第三通 信次数, 且第四通信次数与第三通信次数之间的差值大于预设差值时, 确定任一团体内的成员为核心成员;
当第一团体和第二团体中的任一团体的第三通信次数与第四通信 次数之间的差值 d、于预设差值时, 确定任一团体内的成员为中间人。
需要说明的是: 上述实施例提供的成员角色确定装置在确定成员角 色时, 仅以上述各功能模块的划分进行举例说明, 实际应用中, 可以根 据需要而将上述功能分配由不同的功能模块完成, 即将设备的内部结构 划分成不同的功能模块, 以完成以上描述的全部或者部分功能。 另外, 上述实施例提供的成员角色确定装置与成员角色确定方法实施例属于 同一构思, 其具体实现过程详见方法实施例, 这里不再赘述。
图 7是本申请实施例提供的一种终端的结构示意图。 该终端可以用 功能。 具体来讲:
终端 700可以包括 RF ( Radio Frequency, 射频) 电路 110、 包括有 一个或一个以上计算机可读存储介质的存储器 120、 输入单元 130、 显 示单元 140、 传感器 150、 音频电路 160、 传输模块 170、 包括有一个或 者一个以上处理核心的处理器 180、 以及电源 190等部件。 本领域技术 人员可以理解, 图 7中示出的终端结构并不构成对终端的限定, 可以包 括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。 其中:
RF电路 110可用于收发信息或通话过程中,信号的接收和发送,特 别地, 将基站的下行信息接收后, 交由一个或者一个以上处理器 180处 理; 另外, 将涉及上行的数据发送给基站。 存储器 120可用于存储软件 程序以及模块, 如上述示例性实施例所示出的终端所对应的软件程序以 及模块, 处理器 180通过运行存储在存储器 120的软件程序以及模块, 从而执行各种功能应用以及数据处理, 如实现成员角色确定等。 存储器 120还可以包括存储器控制器, 以提供处理器 180和输入单元 130对存 储器 120的访问。
输入单元 130可用于接收输入的数字或字符信息, 以及产生与用户 设置以及功能控制有关的键盘、 鼠标、 操作杆、 光学或者轨迹球信号输 入。 具体地, 输入单元 130可包括触敏表面 131以及其他输入终端 132。 显示单元 140可用于显示由用户输入的信息或提供给用户的信息以及终 端 700的各种图形用户接口, 这些图形用户接口可以由图形、 文本、 图 标、 视频和其任意组合来构成。
终端 700还可包括至少一种传感器 150, 比如光传感器、 运动传感 器以及其他传感器。 音频电路 160、 扬声器 161, 传声器 162可提供用 户与终端 700之间的音频接口。 终端 700通过传输模块 170可以帮助用 户收发电子邮件、 浏览网页和访问流式媒体等, 它为用户提供了无线或 有线的宽带互联网访问。
处理器 180是终端 700的控制中心, 利用各种接口和线路链接整个 手机的各个部分, 通过运行或执行存储在存储器 120 内的软件程序和 / 或模块, 以及调用存储在存储器 120内的数据, 执行终端 700的各种功 能和处理数据, 从而对手机进行整体监控。 处理器 180可包括一个或多 个处理核心。
终端 700还包括给各个部件供电的电源 190 (比如电池)。
终端 700还包括有存储器以及至少一条指令、 至少一段程序、 代码 集或指令集, 其中至少一条指令、 至少一段程序、 代码集或指令集存储 于存储器中, 且经配置以由一个或者一个以上处理器加载并执行, 以实 根据指定团体的成员通信记录创建虚拟网络;
获取虚拟网络中的核心子网络;
根据核心子网络创建邻接矩阵;
从邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系数, 直至获取到仅包括第一元素和第二元素的指定矩阵;
根据指定矩阵, 将指定团体划分为第一团体和第二团体;
根据第一团体和第二团体的成员的通信行为特征, 确定第一团体和 第二团体的成员角色。
在可能的实现方式中, 上述指令、 程序、 代码集或指令集由一个或 者一个以上处理器加载并执行时, 还可以实现为每个节点添加与本端节 点连接的边的操作, 或者设置邻接矩阵中各个位置上的元素的操作等, 在此不再赘述。
图 8是本申请实施例提供的一种服务器的结构示意图,该服务器 800 可因配置或性能不同而产生比较大的差异, 可以包括一个或一个以上中 央处理器(central processing units, CPU ) 822 (例如, 一个或一个以上 处理器)和存储器 832, 一个或一个以上存储应用程序 842或数据 844 的存储介质 830 (例如一个或一个以上海量存储设备)。 其中, 存储器 832和存储介质 830可以是短暂存储或持久存储。 存储在存储介质 830 的程序可以包括一个或一个以上模块(图示没标出), 每个模块可以包 括对服务器中的一系列指令操作。
服务器 800还可以包括一个或一个以上电源 826, 一个或一个以上 有线或无线网络接口 850, 一个或一个以上输入输出接口 858, 一个或 一个以上键盘 856, 和 /或, 一个或一个以上操作系统 841。
更进一步地, 中央处理器 822可以设置为与存储介质 830通信, 加 载存储介质 830中的一系列指令操作, 并在服务器 800上执行上述成员 角色确定方法, 包括:
根据指定团体的成员通信记录创建虚拟网络;
获取虚拟网络中的核心子网络;
根据核心子网络创建邻接矩阵;
从邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系数, 直至获取到仅包括第一元素和第二元素的指定矩阵;
根据指定矩阵, 将指定团体划分为第一团体和第二团体;
根据第一团体和第二团体的成员的通信行为特征, 确定第一团体和 第二团体的成员角色。
在可能的实现方式中, 中央处理器 822可以加载存储介质 830中的 一系列指令操作, 并在服务器 800上实现为每个节点添加与本端节点连 接的边的操作, 或者设置邻接矩阵中各个位置上的元素的操作等, 在此 不再赘述。
本申请实施例还提供了一种成员角色确定装置, 该成员角色确定装 置包括处理器和存储器,存储器中存储有至少一条指令、至少一段程序、 代码集或指令集, 指令、 程序、 代码集或指令集由处理器加载并执行以 本申请实施例还提供了一种计算机可读存储介质, 该计算机可读存 储介质中存储有至少一条指令、 至少一段程序、 代码集或指令集, 该指 令、 该程序、 该代码集或该指令集由处理器加载并执行以实现上述实施 例的成员角色确定方法中所执行的操作。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤 可以通过硬件来完成, 也可以通过程序来指令相关的硬件完成, 所述的 程序可以存储于一种计算机可读存储介质中, 上述提到的存储介质可以 是只读存储器, 磁盘或光盘等。
以上所述仅为本申请的较佳实施例, 并不用以限制本申请, 凡在本 申请的精神和原则之内, 所作的任何修改、 等同替换、 改进等, 均应包 含在本申请的保护范围之内。

Claims

权利要求书
1、 一种成员角色确定方法, 应用于分析设备, 其特征在于, 所述 方法包括:
根据指定团体的成员通信记录创建虚拟网络, 所述虚拟网络包括多 个节点, 每个节点代表所述指定团体中的一个成员, 通信过的任两个成 员所对应的两个节点连接构成一条边;
获取所述虚拟网络中的核心子网络, 所述核心子网络为所述虚拟网 络中去除度小于预设核数的节点以及所述度小于预设核数的节点连接 的边之后剩余的网络;
根据所述核心子网络创建邻接矩阵, 所述邻接矩阵以所述核心子网 络中的节点作为行和列, 行节点的排列顺序和列节点的排列顺序相同, 且所述邻接矩阵中每个位置上的元素根据对应的行节点与列节点之间 的连接关系确定;
从所述邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系 数, 直至获取到仅包括第一元素和第二元素的指定矩阵, 所述第一维度 为行或者列, 所述第一元素大于所述第二元素;
根据所述指定矩阵, 将所述指定团体划分为第一团体和第二团体, 其中所述指定矩阵中所述第一元素对应的行成员和列成员位于所述第 一团体和所述第二团体中的同一团体, 而所述第二元素对应的行成员和 列成员位于所述第一团体和所述第二团体中的不同团体;
根据所述第一团体和所述第二团体的成员的通信行为特征, 确定所 述第一团体和所述第二团体的成员角色。
2、 根据权利要求 1 所述的方法, 其特征在于, 所述根据所述核心 子网络创建邻接矩阵之前, 所述方法还包括:
在所述核心子网络中, 为每个节点添加与本端节点连接的边, 以保 证根据所述核心子网络创建所述邻接矩阵时, 所述邻接矩阵中对角线上 的元素大于 0。
3、 根据权利要求 1 所述的方法, 其特征在于, 所述根据所述核心 子网络创建邻接矩阵, 包括:
对于所述邻接矩阵中的每个位置, 当所述位置对应的行节点与列节 点连接时, 将所述位置上的元素设置为第三元素;
当所述位置对应的行节点与列节点未连接时, 将所述位置上的元素 设置为第四元素。
4、 根据权利要求 1 所述的方法, 其特征在于, 所述根据所述核心 子网络创建邻接矩阵之前, 所述方法还包括:
将所述核心子网络中的每条边的权重设置为连接的两个节点所对 应的两个成员之间的通信次数或通信参数;
所述根据所述核心子网络创建邻接矩阵, 包括:
将所述邻接矩阵中每个位置上的元素设置为对应的行节点与列节 点连接构成的边的权重。
5、 根据权利要求 1-4任一项所述的方法, 其特征在于, 所述通信行 为特征包括第一通信次数和第二通信次数, 所述第一通信次数为与同一 团体的成员之间的通信次数, 所述第二通信次数为与其他团体的成员之 间的通信次数;
所述根据所述第一团体和所述第二团体的成员的通信行为特征, 确 定所述第一团体和所述第二团体的成员角色, 包括:
当所述第一团体和所述第二团体中任一团体的第一通信次数大于 第一预设次数, 且第二通信次数大于第二预设次数时, 确定所述任一团 体内的成员为核心成员;
当所述第一团体和所述第二团体中任一团体的第一通信次数小于 第二通信次数, 且第二通信次数大于第二预设次数时, 确定所述任一团 体内的成员为联络人; 当所述第一团体和所述第二团体中任一团体的第一通信次数小于 所述第一预设次数, 且第二通信次数大于第二预设次数时, 确定所述任 一团体内的成员为中间人。
6、 根据权利要求 1-4任一项所述的方法, 其特征在于, 所述通信行 为特征包括第三通信次数和第四通信次数, 所述第三通信次数为向其他 团体的成员发送通信消息的次数, 所述第四通信次数为接收其他团体的 成员所发送的通信消息的次数;
所述根据所述第一团体和所述第二团体的成员的通信行为特征, 确 定所述第一团体和所述第二团体的成员角色, 包括:
当所述第一团体和所述第二团体中任一团体的第三通信次数大于 第四通信次数, 且第三通信次数与第四通信次数之间的差值大于预设差 值时, 确定所述任一团体的成员为底层成员;
当所述第一团体和所述第二团体中的任一团体的第四通信次数大 于第三通信次数, 且第四通信次数与第三通信次数之间的差值大于所述 预设差值时, 确定所述任一团体内的成员为核心成员;
当所述第一团体和所述第二团体中的任一团体的第三通信次数与 第四通信次数之间的差值小于所述预设差值时, 确定所述任一团体内的 成员为中间人。
7、 一种成员角色确定装置, 应用于分析设备, 其特征在于, 所述 装置包括处理器和存储器, 所述存储器中存储可被所述处理器执行的指 令, 当执行所述指令时, 所述处理器用于:
根据指定团体的成员通信记录创建虚拟网络, 所述虚拟网络包括多 个节点, 每个节点代表所述指定团体中的一个成员, 通信过的任两个成 员所对应的两个节点连接构成一条边;
获取所述虚拟网络中的核心子网络, 所述核心子网络为所述虚拟网 络中去除度小于预设核数的节点以及所述度小于预设核数的节点连接 的边之后剩余的网络;
居所述核心子网络创建邻接矩阵, 所述邻接矩阵以所述核心子网 络中的节点作为行和列, 行节点的排列顺序和列节点的排列顺序相同, 且所述邻接矩阵中每个位置上的元素根据对应的行节点与列节点之间 的连接关系确定;
从所述邻接矩阵开始, 重复计算第一维度上各个节点之间的相关系 数, 直至获取到仅包括第一元素和第二元素的指定矩阵, 所述第一维度 为行或者列, 所述第一元素大于所述第二元素;
根据所述指定矩阵, 将所述指定团体划分为第一团体和第二团体, 其中所述指定矩阵中所述第一元素对应的行成员和列成员位于所述第 一团体和所述第二团体中的同一团体, 而所述第二元素对应的行成员和 列成员位于所述第一团体和所述第二团体中的不同团体;
根据所述第一团体和所述第二团体的成员的通信行为特征, 确定所 述第一团体和所述第二团体的成员角色。
8、 根据权利要求 7所述的装置, 其特征在于, 当执行所述指令时, 所述处理器进一步用于:
在所述核心子网络中, 为每个节点添加与本端节点连接的边, 以保 证根据所述核心子网络创建所述邻接矩阵时, 所述邻接矩阵中对角线上 的元素大于 0。
9、 根据权利要求 7所述的装置, 其特征在于, 当执行所述指令时, 所述处理器进一步用于:
对于所述邻接矩阵中的每个位置, 当所述位置对应的行节点与列节 点连接时, 将所述位置上的元素设置为第三元素;
当所述位置对应的行节点与列节点未连接时, 将所述位置上的元素 设置为第四元素。
10、根据权利要求 7所述的装置, 其特征在于, 当执行所述指令时, 所述处理器进一步用于:
将所述核心子网络中的每条边的权重设置为连接的两个节点所对 应的两个成员之间的通信次数或通信参数;
将所述邻接矩阵中每个位置上的元素设置为对应的行节点与列节 点连接构成的边的权重。
11、 根据权利要求 7-10任一项所述的装置, 其特征在于, 所述通信 行为特征包括第一通信次数和第二通信次数, 所述第一通信次数为与同 一团体的成员之间的通信次数, 所述第二通信次数为与其他团体的成员 之间的通信次数; 当执行所述指令时, 所述处理器进一步用于:
当所述第一团体和所述第二团体中任一团体的第一通信次数大于 第一预设次数, 且第二通信次数大于第二预设次数时, 确定所述任一团 体内的成员为核心成员;
当所述第一团体和所述第二团体中任一团体的第一通信次数小于 第二通信次数, 且第二通信次数大于第二预设次数时, 确定所述任一团 体内的成员为联络人;
当所述第一团体和所述第二团体中任一团体的第一通信次数小于 所述第一预设次数, 且第二通信次数大于第二预设次数时, 确定所述任 一团体内的成员为中间人。
12、 根据权利要求 7-10任一项所述的装置, 其特征在于, 所述通信 行为特征包括第三通信次数和第四通信次数, 所述第三通信次数为向其 他团体的成员发送通信消息的次数, 所述第四通信次数为接收其他团体 的成员所发送的通信消息的次数; 当执行所述指令时, 所述处理器进一 步用于:
当所述第一团体和所述第二团体中任一团体的第三通信次数大于 第四通信次数, 且第三通信次数与第四通信次数之间的差值大于预设差 值时, 确定所述任一团体的成员为底层成员; 当所述第一团体和所述第二团体中的任一团体的第四通信次数大 于第三通信次数, 且第四通信次数与第三通信次数之间的差值大于所述 预设差值时, 确定所述任一团体内的成员为核心成员;
当所述第一团体和所述第二团体中的任一团体的第三通信次数与 第四通信次数之间的差值小于所述预设差值时, 确定所述任一团体内的 成员为中间人。
13、 一种计算机可读存储介质, 其特征在于, 所述计算机可读存储 介质中存储有至少一条指令、 至少一段程序、 代码集或指令集, 所述指 令、 所述程序、 所述代码集或所述指令集由处理器加载并执行以实现如 权利要求 1 至 6任一权利要求所述的成员角色确定方法中所执行的操 作。
PCT/CN2018/097759 2017-08-30 2018-07-31 成员角色确定方法、装置及存储介质 WO2019042060A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710761789.6A CN110020341B (zh) 2017-08-30 2017-08-30 成员角色确定方法、装置及存储介质
CN201710761789.6 2017-08-30

Publications (1)

Publication Number Publication Date
WO2019042060A1 true WO2019042060A1 (zh) 2019-03-07

Family

ID=65525845

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/097759 WO2019042060A1 (zh) 2017-08-30 2018-07-31 成员角色确定方法、装置及存储介质

Country Status (2)

Country Link
CN (1) CN110020341B (zh)
WO (1) WO2019042060A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113110900B (zh) * 2020-01-10 2023-05-26 钉钉控股(开曼)有限公司 界面显示方法、设备和存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008120072A1 (en) * 2007-04-03 2008-10-09 Fernando Luege Mateos Method and system of classifying, ranking and relating information based on networks
CN101408901A (zh) * 2008-11-26 2009-04-15 东北大学 一种基于主题词条的跨类型数据的概率聚类方法
CN102202012A (zh) * 2011-05-30 2011-09-28 中国人民解放军总参谋部第五十四研究所 通信网络的社团划分方法与系统
CN105608329A (zh) * 2016-01-26 2016-05-25 中国人民解放军国防科学技术大学 一种基于社团演化的组织行为异常检测方法
CN106599273A (zh) * 2016-12-23 2017-04-26 贾志娟 一种基于社交网络分析的微博诈骗团体挖掘方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110035444A1 (en) * 2009-08-06 2011-02-10 Timedright Inc. Relationship security in online social and professional networks and communities
CN105577434B (zh) * 2015-12-21 2019-09-13 中国电子科技集团公司第十五研究所 一种基于社交网络的多重关联挖掘的方法和装置
CN106327345A (zh) * 2016-09-12 2017-01-11 电子科技大学 一种基于多网络模块度的社团发现方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008120072A1 (en) * 2007-04-03 2008-10-09 Fernando Luege Mateos Method and system of classifying, ranking and relating information based on networks
CN101408901A (zh) * 2008-11-26 2009-04-15 东北大学 一种基于主题词条的跨类型数据的概率聚类方法
CN102202012A (zh) * 2011-05-30 2011-09-28 中国人民解放军总参谋部第五十四研究所 通信网络的社团划分方法与系统
CN105608329A (zh) * 2016-01-26 2016-05-25 中国人民解放军国防科学技术大学 一种基于社团演化的组织行为异常检测方法
CN106599273A (zh) * 2016-12-23 2017-04-26 贾志娟 一种基于社交网络分析的微博诈骗团体挖掘方法

Also Published As

Publication number Publication date
CN110020341A (zh) 2019-07-16
CN110020341B (zh) 2022-09-16

Similar Documents

Publication Publication Date Title
CN107102941B (zh) 一种测试用例的生成方法及装置
WO2018177124A1 (zh) 业务处理方法、装置、数据共享系统及存储介质
US10268750B2 (en) Log event summarization for distributed server system
WO2019042180A1 (zh) 资源配置方法及相关产品
WO2021189055A2 (en) Predictive performance analysis for file systems
WO2019042294A1 (zh) 资源配置方法及相关产品
CN107172196B (zh) 一种业务互动方法、装置和系统
CN106375478B (zh) 一种移动终端数据的同步方法、装置及系统
CN106332307A (zh) 一种应用程序接入网络的方法及移动终端
WO2019091028A1 (zh) 应用软件恶意行为的动态告警方法和终端
CN105573733A (zh) 浏览器与web前端通信的方法、web前端及系统
CN105468513A (zh) 一种基于移动终端的测试方法、装置及系统
CN114095567A (zh) 数据访问请求的处理方法、装置、计算机设备及介质
CN107613091A (zh) 一种建立通信的方法及终端
CN103634419A (zh) 终端的远程控制方法及终端
CN108647130A (zh) 一种故障矿机的定位方法、报警方法以及相关设备和系统
CN113034189A (zh) 归因方法、装置、计算机设备和存储介质
CN116431282A (zh) 一种云虚拟主机服务器管理方法、装置、设备及存储介质
CN110554885A (zh) 子应用生成方法、装置、电子设备及存储介质
WO2019042060A1 (zh) 成员角色确定方法、装置及存储介质
CN107332988A (zh) 信息处理方法、移动终端及计算机可读存储介质
CN111930565A (zh) 分布式管理系统中组件的进程故障自愈方法、装置及设备
CN111182522A (zh) 一种团伙确定方法、装置、电子设备及计算机存储介质
CN113891441A (zh) 网络连接方法、装置和电子设备
US9674160B2 (en) Methods for anti-fraud masking of a universal resource indentifier (“URI”)

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18852102

Country of ref document: EP

Kind code of ref document: A1