WO2019192310A1 - 团网络识别方法、装置、计算机设备及计算机可读存储介质 - Google Patents

团网络识别方法、装置、计算机设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2019192310A1
WO2019192310A1 PCT/CN2019/078349 CN2019078349W WO2019192310A1 WO 2019192310 A1 WO2019192310 A1 WO 2019192310A1 CN 2019078349 W CN2019078349 W CN 2019078349W WO 2019192310 A1 WO2019192310 A1 WO 2019192310A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
edge
group
data transfer
attribute value
Prior art date
Application number
PCT/CN2019/078349
Other languages
English (en)
French (fr)
Inventor
夏彬
吴鸣
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
Priority to JP2020545337A priority Critical patent/JP7073514B2/ja
Publication of WO2019192310A1 publication Critical patent/WO2019192310A1/zh
Priority to US16/903,480 priority patent/US10958529B2/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/34—Betting or bookmaking, e.g. Internet betting
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12—Discovery or management of network topologies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/149—Network analysis or design for prediction of maintenance

Definitions

  • the present application relates to the field of data processing, and in particular, to a group network identification method, apparatus, computer device, and computer readable storage medium.
  • the sub-relational network mining in the data transmission relation network is mainly realized by the label diffusion algorithm, which generally has the problem of confusing the data transmission in the group with the target characteristics and the data transmission in the group without the target characteristics.
  • a group network identification method, apparatus, server, and computer readable storage medium are provided.
  • An embodiment of the present application provides a method for identifying a network of a group, including:
  • the computer device confirms the sending direction of the attribute value of each node according to the data transfer feature of the target group to be identified;
  • the computer device in the constructed data transmission relationship network, along the transmitting direction of the first node and the adjacent second node, the edge vector of the edge and the first node attribute value Sending to the second node, the edge vector includes a plurality of feature values of data transfer;
  • the computer device performs weighting calculation on the edge vector received by the second node according to preset calculation logic to obtain an edge of the optimal weight, and the calculation logic matches the data transfer feature of the target group to be identified. ;
  • the computer device confirms each node that is in the target group according to the attribute value of each node after the iteration, and confirms the attributes of each node in the target group.
  • An embodiment of the present application provides a group network identification apparatus, including:
  • a confirmation module configured to confirm the sending direction of the attribute value of each node according to the data transfer feature of the target group to be identified
  • a sending module configured to, in the constructed data transmission network, the side vector of the edge and the first node attribute along an edge of the first node connected to the adjacent second node along the sending direction Sending a value to the second node, the edge vector including a plurality of feature values of the data transfer;
  • a calculation module configured to perform weighting calculation on an edge vector received by the second node according to preset calculation logic to obtain an edge of an optimal weight, where the calculation logic is related to the data transfer feature of the target group to be identified match;
  • An update module configured to update an attribute value of the second node according to an attribute value of the first node that is connected by an edge of the optimal weight
  • An iterative module configured to trigger an iterative execution to send the edge vector of the edge and the first node attribute value to the edge of the first node connected to the adjacent second node along the sending direction a second node, the edge vector received by the second node is weighted according to a preset calculation logic, and the obtained edge of the optimal weight, and the first node connected according to the edge of the optimal weight
  • the attribute value the step of updating the attribute value of the second node until the iteration meets the preset stop condition
  • the confirmation module is further configured to confirm each node of the target group according to an attribute value of each node after the iteration, and confirm an attribute of each node in the target group.
  • An embodiment of the present application provides, in an aspect, a computer device, including a memory and a processor, where the computer stores computer readable instructions, when the computer readable instructions are executed by the processor, causing the processor to execute The group network identification method provided by the embodiment of the present application.
  • Embodiments of the present application provide, in an aspect, one or more non-volatile storage media storing computer readable instructions, when executed by one or more processors, causing one or more processors to execute The group network identification method provided by the embodiment of the present application.
  • FIG. 1 is an application environment diagram of a group network identification method according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of a data transmission relationship network in a group network identification method according to an embodiment of the present disclosure
  • FIG. 3 is a flowchart of a method for identifying a group network according to an embodiment of the present application
  • FIG. 4 is a flowchart of a method for identifying a group network according to an embodiment of the present application
  • FIG. 5 is a schematic diagram of a virtual resource transfer relationship network in a group network identification method according to an embodiment of the present disclosure
  • FIG. 6 is a schematic diagram of the virtual resource transfer relationship network in FIG. 5 after the initial cluster core index is allocated;
  • FIG. 7 is a schematic diagram of a virtual resource transfer relationship network in FIG. 6 when transmitting a multi-dimensional vector
  • FIG. 8 is a schematic diagram of a target group identified from the virtual resource transfer relationship network in FIG. 7;
  • FIG. 9 is a schematic structural diagram of a group network identification apparatus according to an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of a group network identification apparatus according to an embodiment of the present application.
  • FIG. 11 is a schematic structural diagram of hardware of a computer device according to an embodiment of the present application.
  • FIG. 1 is an application environment diagram of a group network identification method according to an embodiment of the present application.
  • a plurality of terminals 100 constitute a data transmission relationship network
  • the storage server 200 records related information data for data exchange between the terminals 100 in the data transmission relationship network
  • the server 300 records related information data according to the storage server 200.
  • the target group is identified from the data transmission network by the group network identification method provided in the following embodiments.
  • server 300 can be a distributed server cluster of pregel graph computing frameworks.
  • Server 200 can also be a distributed server cluster.
  • the pregel graph calculation framework is the computational framework of distributed graph computation.
  • the graph consists of nodes and edges.
  • the data transmission relationship network refers to a network topology composed of a relationship between at least two nodes and each node.
  • the three computer devices are abstracted into three nodes: node A, node B, and node C, and The side connecting the two nodes represents the communication relationship between the two computer devices, thus forming a data transmission relationship network composed of these three computer devices.
  • the storage server records related information data, such as the communication duration of one computer device with another computer device, the communication frequency of one computer device with another computer device during a preset time period, and the communication of one computer device with another computer device Time period.
  • a group refers to a group of nodes having the same characteristics in a data transmission network.
  • a node has a data transfer relationship between nodes.
  • a group network is a network topology formed by all data nodes in a group due to data transmission relationships.
  • the node in a data transmission relationship network having a virtual resource transfer relationship, the node is a computer device corresponding to an account for transferring a virtual resource, and when the virtual resource is transferred between the nodes, the virtual resource transfer keyword of the remark is “admission, gambling” Words related to gambling, such as "chips", correspondingly, these nodes constitute a gambling-like group.
  • the virtual resource transfer relationship between these nodes and nodes constitutes a group network.
  • FIG. 2 is a schematic diagram of a data transmission relationship network in a group network identification method according to an embodiment of the present application.
  • the data transmission relationship network in the embodiments of the present application may specifically be a virtual resource transfer relationship network, but the virtual resource transfer relationship shall not constitute any limitation on the type of the data transmission relationship network to which the present application is applicable.
  • the virtual resource transfer relationship network includes a plurality of nodes, and the node represents a virtual resource transfer party or a computer device corresponding to the receiver, such as a computer device used by a fund transfer account or a fund receiving account, and each side represents two connected terminals. There is data transmission between nodes, that is, there is a virtual resource transfer relationship between the two nodes to form an edge.
  • the virtual resource may refer to a resource that can be used for redemption of a product or a service, such as a currency, a point, a golden bean, a gift voucher, a voucher, a coupon, a gift card, a wealth management fund, and the like, which are not specifically limited in the embodiment of the present invention.
  • Virtual resources can be transferred on the Internet platform.
  • FIG. 2 is a virtual resource transfer relationship network
  • the nodes in the virtual resource transfer relationship network include: node A, node B, and node C.
  • the direction of the arrow between the nodes indicates the virtual resource transfer relationship, that is, the node B performs virtual resource transfer to the node A, and the node C performs virtual resource transfer to the node B.
  • the group needs to be divided from a large number of nodes according to the characteristics of the target group to be identified.
  • the target group may be in the virtual resource transfer relationship network, belonging to the gambling gang.
  • the attribute can refer to the core level of the node in the group.
  • the cluster network identification method, apparatus, computer apparatus, and computer readable storage medium are described in detail below in various embodiments.
  • the pregel graph calculation framework and the api in the spark graphx can be used for group network identification, and data transmission information of each node in the data transmission relationship network is stored in multiple servers, and the implementation is performed according to the data transmission information.
  • the network identification method in the group is described in detail below in various embodiments.
  • FIG. 3 is a flowchart of a method for identifying a group network according to an embodiment of the present application. As shown in FIG. 3, the method is performed by a server, and the method includes:
  • the server confirms the sending direction of the attribute values of each node according to the data transfer characteristics of the target group to be identified.
  • the server confirms that the attribute value sending direction of each node is the reverse direction of the data transfer between the nodes.
  • the divergence type is that there are multiple data receiver nodes for the same data sender node, for example, node A transfers data to node B and node C, node B transfers data to node D and node E, and node C transfers data to node F and Node G, in the group consisting of ABCDEFG, the data is transferred from one node A through node B and node C to four nodes D, E, F, G, and the data transfer feature is divergent; the above-mentioned divergent target group
  • the same data sender node will have at least 2 data receiver nodes corresponding to each other.
  • the server confirms that the attribute value of each node is sent in the positive direction of data transfer between the nodes, and the convergence type is that the pointer has multiple data on the same data receiver node.
  • the sender node for example, node B and node C transfer data to node A, node D and node E transfer data to node B, node F and node G transfer data to node C, then in the group consisting of ABCDEFG, data slave node
  • the four nodes D, E, F, and G are transferred to a node A through the node B and the node C, and the data transfer feature is a convergent type; the same data receiver node of the convergent target group has at least two data transmissions correspondingly. Square node.
  • the data transmission relationship network is a passenger transportation network
  • the node is a computer equipment corresponding to the docking station.
  • the target group to be identified is the to-be-opened station group composed of the docking station with the most value of the passenger line, and the station group to be opened is to be opened.
  • the characteristics of the transportation data transfer are convergence type, that is, when the passenger vehicles are all gathered to one site, when identifying the network formed by the to-be-opened site group, it is confirmed that the attribute value of each node is sent in the direction of data transfer between the nodes. direction.
  • the edge vector of the edge and the first node attribute value are sent to the second node along the sending direction on the edge where the first node is connected to the adjacent second node.
  • the data transmission relationship network may be constructed according to the information data recorded in the storage server 200 in FIG. 1, and the constructed data transmission relationship network corresponds to the actual existing data transmission network.
  • the edge vector includes a plurality of feature values of the data transition, that is, the edge vector is used to describe the characteristics of the data transition according to the at least one single-dimensional vector, for example, the edge vector may include a key in the data transfer feature of the target group to be identified. The number of times the word is matched, the single-dimensional vector such as the frequency of data transfer, the number of transfers, and the total number of transfers in the preset period.
  • the keyword refers to a keyword corresponding to the feature of the target group to be identified.
  • the resource transfer keyword is a word related to gambling.
  • the server in the passenger transportation network, can collect passenger transportation relationship data for each month of each year within one year with a sampling period of one month.
  • the edge vector may include the average number of passengers transported between stations in a year, the number of passengers transported between stations during each year within a month, and the variance of the number of passengers transported between stations during each year. .
  • An edge is an edge formed between nodes due to a data transfer relationship between nodes. For example, if the A node transfers data to the B node, an edge is formed between the A node and the B node.
  • the attribute value refers to the parameter value that characterizes the node attribute, and may include the node ID (Identification) and the cluster core index of the node and other parameter values that characterize the node attribute.
  • the group core index can represent the core degree of a node in the data transmission relationship network.
  • the second node may simultaneously transfer data to multiple first nodes, or may be multiple first nodes to simultaneously transfer data to the second node, so along the sending In the direction, a plurality of edge vectors and attribute values of the first node are sent to the second node, the second node may be a transfer party or a transfer party of the data transfer, and the second node is a receiver of the attribute value.
  • the specific form of sending can be sent by message, which contains edge vectors and attribute values.
  • the edge vector is sent along the transmission direction. If the target group is divergent, the attribute of the data transfer destination is used to affect the data transfer. Attribute; if the target group is convergent, the attributes of the transferee of the data transfer are used to affect the attributes of the transferee of the data transfer.
  • the algorithm of the embodiment of the present application may be completed in the server, so the server may send the edge vector of the edge and the first node attribute value to the second node to start the calculation process of the target group.
  • the edge of the optimal weight is the edge with the largest weight between the first node and the second node, and the edge represents the largest weight of the data transfer relationship.
  • the calculation logic is a weight-dependent calculation logic that matches the data transfer characteristics of the target group to be identified, that is, the edge of the optimal weight differs depending on the calculation logic, and the weight is matched with the data transfer characteristic of the target group. .
  • the weight value of the feature value of each data transfer in the edge vector is given by the calculation logic, and the weight value assigned to the feature value of the data transfer by the calculation logic is matched by the data transfer feature of the target group.
  • the confirmation of the edge of the optimal weight has a significant impact on the final confirmed group.
  • the weight values different in the feature values of the data transitions are given in advance, for example, the data transfer frequency, the number of transitions, and the number of transitions are respectively assigned weight values of 0.6, 0.3, and 0.1, and then, on the first side,
  • the edge vector includes the data transfer frequency is 1 time per day, the number of transfers is 4 times, the number of transfers is 1 million (in units of 100,000, which is equivalent to 10); on the second side, the edge vector includes the data transfer frequency is 1.5 per day.
  • the number of transfers is 6 times, the number of transfers is 200,000 (in units of 100,000, which is equivalent to 2), and the weighting calculation is performed separately.
  • each second node has one and only one first node is connected to the second node by using an edge of the optimal weight, and the second node is updated according to the attribute value of the first node.
  • the attribute value of the node is the attribute value of the node.
  • the iterative execution sends the edge vector of the edge and the first node attribute value to the second node along the sending direction of the attribute value on the edge of the first node and the adjacent second node, and
  • the plurality of edge vectors received by the second node perform weighting calculation according to preset calculation logic, obtain the edge of the optimal weight, and update the attribute of the second node according to the attribute value of the first node connected by the edge of the optimal weight
  • the step of value that is, iteratively performing steps 101-104, continuously updates the attribute value of the side of the receiving edge vector until the iteration conforms to the preset stop condition.
  • the preset stop condition may be that the number of iterations exceeds a preset number of iterations thresholds, or the time of the iterative calculation exceeds a preset iteration time threshold.
  • the preset stop condition may also be in the data transmission network, without the second node that can update the attribute value.
  • step 107 is performed; otherwise, the next iteration is continued.
  • the attribute value of the node can identify the nodes belonging to the same group, and thus the composition of the target group network.
  • the attribute value of the node can also confirm the attributes of each node in the group.
  • the attribute refers to the core degree of each node in the group, that is, the importance degree of each node.
  • attributes such as suppliers, agents, and customers in the marketing group can be confirmed according to the attribute values of the nodes.
  • the attributes of the dealer, the claim adjuster, the gambler and the like in the gambling group can be confirmed according to the attribute value of the node.
  • the group network identification method provided by the embodiment sends the edge vector of the edge between the first node and the second node to the second node by using the sending direction confirmed according to the data transfer feature of the target group to be identified, and Receiving the edge vector for the second node, performing weighting calculation according to the preset calculation logic, obtaining the edge of the optimal weight, updating the attribute value of the second node according to the attribute value of the first node connected by the edge of the optimal weight, and iterating multiple times After the above steps, the attribute of the node and the node in the graph are confirmed according to the attribute value of the node after the iteration.
  • the edge vector includes a plurality of feature values describing the data transfer feature between the second node and the first node, the accuracy can be more accurately
  • the target group having the data transfer feature is distinguished, and since the iteration can be limited to updating the attribute value of the second node with the attribute value of the first node, such one-way calculation reduces the time complexity of the iteration.
  • FIG. 4 is a flowchart of a method for identifying a network of a group according to another embodiment of the present application. As shown in the figure, the method includes:
  • the server initializes the attribute value of each node in the data transmission relationship network, and the attribute values include: a node ID (Identity) and a group core index. Then, the attribute value of each node is initialized as: the node ID of the first node is used as the initial ID of the first node, the node ID of the second node is used as the initial ID of the second node, and the first The cluster core index of the node and the second node are both preset values, and optionally, the preset value is 0.
  • the group core index can be used to measure the importance of each node in the data transmission network.
  • the update rule of the group core index may be preset, and may include adding or decreasing an update value each time based on the preset value, and the update value may also be customized, for example, the update value is 1, and the update value is 10. Etc., the updated value can be any value. As long as the same update value is set in the entire network of the data transmission network, the importance level of determining a node is not affected.
  • the update rule of the cluster core index is to add an update value each time based on the preset value, the smaller the cluster core index of a node is, the node is in the core position in the cluster;
  • the core index is updated by reducing an update value each time based on the preset value, and the node core index of a node is larger, and the node is at the core position in the cluster.
  • the server confirms the sending direction of the attribute values of each node according to the data transfer characteristics of the target group to be identified.
  • the attribute value sending direction of each node is the reverse direction of data transfer between the nodes.
  • the attribute value transmission direction of each node is the positive direction of data transfer between the nodes.
  • the edge vector of the edge and the first node attribute value are sent to the second node along the sending direction on the edge where the first node is connected to the adjacent second node;
  • the server sends the edge vector of the edge and the first node attribute value to the second node along the sending direction on the edge of the first node connected to the adjacent second node.
  • the edge vector includes a plurality of feature values of the data transfer, and specifically includes: a number of times the keyword of the data transfer matches the keyword in the data transfer feature of the target group, a frequency of transfer of the data in the preset period, a number of transfers, and a number of transfers.
  • the number of times the keyword of the data transfer matches the keyword in the data transfer feature of the target group means that there are many keywords in the data transfer feature of the target group, and the additional information in the data transfer matches the keywords.
  • the number of times can indicate whether the target group is the same important characteristic value.
  • the target group is a marketing group.
  • the key words in the data transfer characteristics can include: sales money, agency fee, amount, various cargo names, and The more times the keywords in the additional information for data transfer between the nodes match the keywords, the more likely the node belongs to the marketing group, and the additional information may be notes, messages, and the like.
  • the feature value of the data transfer in the plurality of edge vectors received by the second node is the same type of feature value according to the weight of each feature value in the preset calculation logic. The values are compared, and the edge corresponding to the feature value with the largest value is the edge of the optimal weight.
  • the data transmission relationship network is a virtual resource transfer relationship network
  • the edge vector received by the second node is a feature vector of the virtual resource transfer
  • the feature of the virtual resource transfer in the multiple edge vectors received by the second node is The value is compared with the value of the feature value of each type according to the weight of each feature value in the preset calculation logic, and the edge corresponding to the feature value with the largest value is the edge of the optimal weight.
  • the feature values of the plurality of data transfers include the number of times the keyword of the data transfer matches the keyword in the data transfer feature of the target group, the frequency of transfer of the data in the preset period, the number of transfers, and the number of transfers.
  • the feature value of the data transfer in the edge vector received by the second node is the same type of weight according to the weight of each of the feature values in the preset calculation logic.
  • the values of the values are compared, and the edges with the largest value of the eigenvalue corresponding to the optimal weight include:
  • the number of times the keyword transferred by the data is matched with the keyword in the data transfer feature of the target group, and the transfer frequency, the number of transitions, and the number of transfers of the data transfer in the preset period are respectively according to weights corresponding to the preset calculation logic
  • the values are compared in descending order, and the edge corresponding to the feature value having the largest value is the edge of the optimal weight.
  • the feature value of the multiple virtual resource transitions includes the number of times the keyword of the virtual resource transfer matches the keyword in the virtual resource transfer feature of the target group, the frequency of the virtual resource transfer, the number of transfers, and the transfer amount in the preset period.
  • the edge vector received by the second node is weighted according to a preset calculation logic, and the edge of the optimal weight is obtained, which is the number of times the keyword transferred by the virtual resource matches the keyword in the virtual resource transfer feature of the target group.
  • the transition frequency, the number of transitions, and the transfer amount of the virtual resources in the preset period are compared according to the weights of the respective corresponding weights in the preset calculation logic, and the edges corresponding to the feature values having the largest value are the optimal weights. The side.
  • the edge of the optimal weight is the edge with the largest weight between the first node and the second node, and the virtual resource transfer relationship weight represented by the edge is the largest.
  • the calculation logic is a weight-related calculation logic that matches the virtual resource transfer feature of the target group to be identified, that is, the edge of the optimal weight is different due to different calculation logic, and the weight is virtual with the target group.
  • the resource transfer characteristics are matched.
  • the weight values of the vectors in the edge vector are assigned by the calculation logic, and the weight values assigned to the vectors by the calculation logic are matched by the virtual resource transfer characteristics of the target group.
  • the confirmation of the edge of the optimal weight has a significant impact on the final confirmed group.
  • the preset calculation logic matches the characteristics of the target group.
  • the preset calculation logic is different depending on the characteristics of the group to be confirmed.
  • the node ID of the second node is updated to the node ID of the first node connected by the edge of the optimal weight
  • the node core index of the second node is updated to the edge of the optimal weight according to the update rule.
  • the group core index of the first node increases or decreases by a preset update value, and the update value may be a preset arbitrary value, for example, 1, the update rule may be to update the group core index of the second node to the optimal weight.
  • the group core index of the first node connected to the edge is incremented by 1, or the group core index of the second node is updated to the edge of the first node of the edge of the optimal weight minus one.
  • each node's core index increases by 1, and the node ID of the node is simultaneously Updating to the node ID of other nodes indicates that the importance of the node is not as important as this other node, so the larger the core index of a node, the lower the core level of the node.
  • the core index of the group increased and the core level declined.
  • step 208 is performed; otherwise, the next iteration is continued.
  • the preset stop condition is that the number of iterations of performing step 202 to step 205 reaches a preset number of iterations, or the node ID of each node in the data transmission relationship network does not change after performing step 202 to step 205.
  • the node ID in the attribute value of the first node of the edge of the optimal weight is the same as the node ID of the second node, and the node ID is not updated, and the group core index is not updated.
  • step 208 is performed.
  • the attribute value of the node can identify the nodes belonging to the same group, and thus the composition of the target group network. Specifically, the nodes with the same node ID are the same group.
  • the attribute value of the node can also confirm the attributes of each node in the group, and specifically confirm the attributes of each node according to the update rule and the group core index of each node, and the attribute refers to the core degree of each node in the group, that is, each The importance of the node.
  • the specific confirmation method refer to the related description of the foregoing content, and details are not described herein again.
  • a group network diagram is generated, and nodes of different group core indexes in the group network diagram are marked with different preset distinguishing features, for example, different preset colors are marked. And output the network diagram of the group, and display the nodes according to the visualization of the core index of the cluster, so that the user can visually check the core degree of the nodes of the group.
  • the node ID and the cluster core index of the node are marked on each node in the group network diagram, and the edge vector is marked on the edge between each two connected nodes, which is convenient for the user.
  • the network diagram of the group it is possible to clearly know the information of each node of the network diagram of the group and the relationship between the nodes, provide necessary information for judging the nature of the group, and improve the accuracy of the judgment.
  • the following is an example of mining a gambling gang relationship network from a virtual resource transfer relationship network, and is not limited to any form of the group network identification method provided by this embodiment. . After dig out the gambling gang and report it to the relevant departments, monitoring the gambling gang can be cracked down.
  • FIG. 5 is a schematic diagram of a virtual resource transfer relationship network in a group network identification method according to an embodiment of the present disclosure.
  • the virtual resource transfer relationship network includes 13 nodes: node A, node B, ..., node M.
  • each edge can pass an edge vector
  • each edge vector includes three virtual resource transfer feature values: feature value a1, feature value a2 and feature value a3, these three virtual resource transfer
  • feature values are the number of keyword matching in the virtual resource transfer feature of the keyword and the target group, the number of virtual resource transfers in one week, and the total amount of virtual resources transferred in one week.
  • the feature value a1 may be specifically the number of times the keyword in the note in the note when the virtual resource is transferred matches the gambling-related keyword, such keywords may be, for example, It is: gambling-related words such as "gambling", “chips”, and "big”.
  • the node ID of each node is the initial node ID of each node, and an identical initial cluster core index is defined for each node, and the initial cluster core index may be 0.
  • the initial node ID and the initial cluster core index are all marked next to the corresponding node, as shown in FIG. 6 , the initial node ID of the node A is 1, and the initial cluster core index is 0, and the (1, 0)
  • the form is marked next to node A; as shown in Figure 6, the initial node ID of node B is 2, and the initial group core index is 0, then it is marked in the form of (2,0) next to node B, and the other nodes are in the same form. Label.
  • FIG. 6 is a schematic diagram of the virtual resource transfer relationship network in FIG. 5 after the initial cluster core index is allocated.
  • the initial node IDs of each node in FIG. 6 are different, specifically, in this example. In the middle, the initial cluster core index of each node is 0, and the update rule of the cluster core index is increased by 1 for each update.
  • the first node is in the form of a message in the opposite direction of the virtual resource transfer direction on each side of the first node and the second node.
  • the two nodes send the edge vector and the attribute value of the first node, where the first node is the receiver of the virtual resource transfer, and the second node is the sender of the virtual resource transfer, see FIG. 7, and FIG. 7 is the virtual in FIG. Resource transfer relationship network, a schematic diagram when sending multidimensional vectors.
  • the edge vector of the plurality of messages received by the second node is weighted according to preset calculation logic to obtain a target message.
  • the node D receives the first message Msg1 and the second message Msg2.
  • the first message Msg1 includes: a node ID of the node A, a group core index of the node A, and a first side vector edge1 connecting the nodes A and the side of the node D.
  • the first edge vector egde1 includes three virtual resource transfer feature values: Msg1.a1, Msg1.a2, and Msg1.a3, which respectively represent the virtual keyword and the target group in the remarks when the node D performs the virtual resource transfer to the node A.
  • the second message Msg2 includes: a node ID of the node B, a group core index of the node B, and a second edge vector edge2, and the second edge vector edge2 includes the feature values of the three virtual resource transitions: Msg2.a1, Msg2.a2, and Msg2.
  • .a3 indicates the number of matches between the keyword and the virtual resource transfer keyword of the target group when the node D performs the virtual resource transfer to the node B, and the number of times the transfer node D transfers the virtual resource to the node B within one week and the node within one week The total amount of virtual resources that D transfers to Node B.
  • the calculation logic for weighting the eigenvalues of each virtual resource transfer is determined, and the eigenvalues of the virtual resource transfer with larger weights are preferentially compared, and vice versa, the virtual resources are transferred with smaller weights.
  • the eigenvalues are compared later.
  • the value of a1 with the largest weight is first compared, that is, the value of Msg1.a1 and the value of Msg2.a1 are compared. If the value of Msg1.a1 is greater than the value of Msg2.a1.
  • the edge of the first message Msg1 is sent as the current optimal weight edge; if the value of Msg1.a1 is smaller than Msg2 The value of .a1, and the absolute value of the difference between the value of Msg1.a1 and the value of Msg2.a1 is greater than the first preset deviation value, the edge of the second message Msg2 is sent as the current optimal weight edge; if Msg1.
  • the comparison weight is smaller than the value of a2 of a1, that is, the value of Msg1.a2 and the value of Msg2.a2 are compared.
  • the value of Msg1.a2 is greater than the value of Msg2.a2, and the absolute value of the difference between the value of Msg1.a2 and the value of Msg2.a2
  • the edge of the first message Msg1 is sent as the current optimal weight edge, wherein the value of Msg1.a2 is greater than the preset value indicating Msg1.
  • A2 constitutes the characteristic value of the meaningful virtual resource transfer; if the value of Msg1.a2 is less than the value of Msg2.a2, and the absolute value of the difference between the value of Msg1.a2 and the value of Msg2.a2 is greater than the second preset deviation value And the value of Msg2.a2 is greater than the preset value, the edge of the second message Msg2 is sent as the current optimal weight edge; if the absolute value of the difference between the value of Msg1.a2 and the value of Msg2.a2 is less than the second The preset deviation value, or the value of Msg1.a2 and the value of Msg2.a2 are not greater than the preset value, and the comparison weight is less than a1 and less than the value of a3 of a2, that is, comparing the value of Msg1.a3 with the value of Msg2.a3 The value of the value, if the value of Msg1.a3 is greater than the value of Msg2.a3,
  • the edge is the current optimal weight edge. If the value of Msg1.a3 is less than the value of Msg2.a3, and the absolute value between the value of Msg1.a3 and the value of Msg2.a3 is greater than the third preset bias Value, then the edge of the second message sends Msg2 as a current optimal weight side.
  • the edge that obtains the optimal weight is the edge of the second message Msg2 sent between the node B and the node D, and the node ID of the node B is updated to the node ID of the node D, and the node B is After the group core index is increased by 1, it is updated to the group core index of node D.
  • FIG. 8 is a schematic diagram of a target group identified from the virtual resource transfer relationship network in FIG.
  • the network of the group is a virtual resource transfer network for a gambling gang.
  • the role of these nodes in the gambling gang can be judged. The smaller the group core index, the more important the node is in the gambling gang.
  • node H, node F, The node core index of node G and node I is 2, which is the peripheral gambler.
  • the node core index of node E and node D is 1, which is the claim adjuster, and the node core index of node B is 0, which is the dealer.
  • all the virtual resource transfer behaviors of the node B may be summarized and reported to the anti-money laundering department.
  • the edge vector of the edge between the first node and the second node is sent to the second node along the sending direction confirmed according to the data transfer feature of the target group to be identified, and the second node is Receiving the edge vector, performing weighting calculation according to the preset calculation logic, obtaining the edge of the optimal weight, updating the attribute value of the second node according to the attribute value of the first node connected by the edge of the optimal weight, and iterating the above steps multiple times according to the step
  • the property value of the node after iteration confirms the group and the attribute of the node in the graph. Since the edge vector includes a plurality of feature values describing the data transfer feature between the second node and the first node, the relationship can be more accurately distinguished.
  • the target group of data transfer features, and since the iteration can be limited to updating the attribute values of the second node with the attribute values of the first node, such one-way computation reduces the time complexity of the iteration.
  • FIG. 9 is a schematic structural diagram of a network identification device according to an embodiment of the present disclosure. As shown in FIG. 9, the network identification device includes:
  • the confirmation module 401 is configured to confirm the sending direction of the attribute value of each node according to the data transfer feature of the target group to be identified;
  • the confirmation module 401 is further configured to: if the data transfer feature of the target group to be identified is a divergent type, confirm that the attribute value sending direction of each node is a reverse direction of data transfer between the nodes;
  • the confirmation module 401 is further configured to: if the data transfer feature of the target group to be identified is a convergence type, confirm that the attribute value transmission direction of each node is the positive direction of data transfer between the nodes.
  • the sending module 402 is configured to, in the constructed data transmission relationship network, the edge vector of the edge and the first node attribute value along the sending direction on the edge where the first node is connected to the adjacent second node.
  • the edge vector includes a plurality of feature values of the data transfer;
  • the calculation module 403 is configured to perform weighting calculation on the edge vector received by the second node according to preset calculation logic to obtain an edge of the optimal weight, and the calculation logic matches the data transfer feature of the target group to be identified;
  • the updating module 404 is configured to update an attribute value of the second node according to an attribute value of the first node connected by the edge of the optimal weight;
  • An iteration module 405, configured to trigger an iterative execution on the edge of the first node connected with the adjacent second node along the sending direction, and send the edge vector and the first node attribute value to the second node, And the edge vector received by the second node is weighted according to a preset calculation logic, and the obtained edge of the optimal weight, and the attribute value of the first node connected according to the edge of the optimal weight, is updated.
  • the confirmation module 401 is further configured to confirm each node of the target group according to the attribute value of each node after the iteration, and confirm the attributes of each node in the target group.
  • the edge vector of the edge between the first node and the second node is sent to the second node along the sending direction confirmed according to the data transfer feature of the target group to be identified, and the second node is Receiving the edge vector, performing weighting calculation according to the preset calculation logic, obtaining the edge of the optimal weight, updating the attribute value of the second node according to the attribute value of the first node connected by the edge of the optimal weight, and iterating the above steps multiple times according to the step
  • the property value of the node after iteration confirms the group and the attribute of the node in the graph. Since the edge vector includes a plurality of feature values describing the data transfer feature between the second node and the first node, the relationship can be more accurately distinguished. This target group of data transfer features, and since the iteration is limited to updating the attribute values of the second node with the attribute values of the first node, this one-way calculation reduces the time complexity of the iteration.
  • FIG. 10 is a schematic structural diagram of a group network identification apparatus according to an embodiment of the present invention. Unlike the group network identification apparatus shown in FIG. 9, in this embodiment:
  • the attribute values include: a node ID and a group core index, and the device further includes:
  • the initialization module 501 is configured to use the node ID of the first node as an initial ID of the first node, the node ID of the second node as an initial ID of the second node, and initialize the first node and the second
  • the node core index of the node is the same as the preset value.
  • the update module 404 is further configured to update the node ID of the second node to the node ID of the first node of the edge of the optimal weight, and update the group core index of the second node to the update rule according to the update rule.
  • the cluster core index of the first node of the edge of the optimal weight is increased or decreased by a preset update value.
  • the confirmation module 401 is further configured to confirm the attributes of the nodes according to the update rule and the group core index of each node.
  • the calculation module 403 is further configured to: in the order of the weight of each of the plurality of edge vectors received by the second node, the weight corresponding to each of the feature values in the preset calculation logic is in descending order The value of the feature value is compared with the same type, and the edge corresponding to the feature value with the largest value is the edge of the optimal weight.
  • the feature value of the plurality of data transitions includes the number of times the keyword of the data transfer matches the keyword in the data transfer feature of the target group, the frequency of transfer of the data in the preset period, the number of transfers, and the amount of transfer, and the calculation module 403 further
  • the number of times the keyword used for transferring the data matches the keyword in the data transfer feature of the target group, and the transfer frequency, the number of transfers, and the transfer amount of the data transfer in the preset period are respectively corresponding to each of the preset calculation logics.
  • the weights of the weights are compared in descending order, and the edges corresponding to the largest eigenvalues are the edges of the optimal weight.
  • the device further includes: a generating module 502, configured to generate a group network map according to a connection relationship of nodes belonging to the same group in the data transmission relationship network; and an annotation module 503, configured to use different groups in the network diagram of the group a node of the core index, labeled with different preset distinguishing features; an output module 504, configured to output the network diagram of the group; and an annotation module 503, configured to label the node ID of the node on each node in the network diagram of the group and the node The core index, and the edge vector is marked on the edge between every two connected nodes.
  • a generating module 502 configured to generate a group network map according to a connection relationship of nodes belonging to the same group in the data transmission relationship network
  • an annotation module 503 configured to use different groups in the network diagram of the group a node of the core index, labeled with different preset distinguishing features
  • an output module 504 configured to output the network diagram of the group
  • an annotation module 503, configured to label the node ID of the no
  • the edge vector of the edge between the first node and the second node is sent to the second node along the sending direction confirmed according to the data transfer feature of the target group to be identified, and the second node is Receiving the edge vector, performing weighting calculation according to the preset calculation logic, obtaining the edge of the optimal weight, updating the attribute value of the second node according to the attribute value of the first node connected by the edge of the optimal weight, and iterating the above steps multiple times according to the step
  • the property value of the node after iteration confirms the group and the attribute of the node in the graph. Since the edge vector includes a plurality of feature values describing the data transfer feature between the second node and the first node, the relationship can be more accurately distinguished. This target group of data transfer features, and since the iteration is limited to updating the attribute values of the second node with the attribute values of the first node, this one-way calculation reduces the time complexity of the iteration.
  • Figure 11 is a diagram showing the internal structure of a computer device in one embodiment.
  • the computer device may specifically be the server 120 of FIG.
  • the computer device includes a processor, a memory, and a network interface connected by a system bus.
  • the memory comprises a non-volatile storage medium and an internal memory.
  • the non-volatile storage medium of the computer device stores an operating system, and can also store computer readable instructions that, when executed by the processor, cause the processor to implement a group network identification method.
  • the internal memory can also store computer readable instructions that, when executed by the processor, cause the processor to perform a network identification method. It will be understood by those skilled in the art that the structure shown in FIG.
  • FIG. 11 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation of the computer device to which the solution of the present application is applied.
  • the specific computer device may include More or fewer components than shown in the figure, or some components combined, or with different component arrangements.
  • the group network identification device provided herein can be implemented in the form of a computer readable instruction that can be executed on a computer device as shown in FIG.
  • the program modules constituting the group network identification device may be stored in a memory of the computer device, such as the confirmation module 401, the transmission module 402, the calculation module 403, the update module 404, the iteration module 405, and the encoding module 912 shown in FIG.
  • the computer readable instructions formed by the various program modules cause the processor to perform the steps in the group network identification method of the various embodiments of the present application described in this specification.
  • the embodiment of the present application further provides a computer readable storage medium, which may be disposed in the group network identification device in the foregoing embodiments.
  • the computer readable storage medium may be the memory in the aforementioned embodiment shown in FIG.
  • the computer readable storage medium stores a computer program that, when executed by the processor, implements the group network identification method described in the foregoing embodiments shown in FIGS. 3 to 8.
  • the computer storable medium may also be a USB flash drive, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
  • the medium of the program code may be disposed in the group network identification device in the foregoing embodiments.
  • the computer readable storage medium may be the memory in the aforementioned embodiment shown in FIG.
  • the computer readable storage medium stores a computer program that, when executed by the processor, implements the group network identification method described in the foregoing embodiments shown in FIGS. 3 to 8.
  • the computer storable medium may also
  • Non-volatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory can include random access memory (RAM) or external cache memory.
  • RAM is available in a variety of formats, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronization chain.
  • Synchlink DRAM SLDRAM
  • Memory Bus Radbus
  • RDRAM Direct Memory Bus Dynamic RAM
  • RDRAM Memory Bus Dynamic RAM
  • the modules described as separate components may or may not be physically separated.
  • the components displayed as modules may or may not be physical modules, that is, may be located in one place, or may be distributed to multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Algebra (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请实施例公开了一种团网络识别方法、装置、服务器及计算机可读存储介质，其中方法包括据待识别的目标团的数据转移特征，确认各节点的属性值发送方向，在数据传输关系网的第一节点与相邻的第二节点连接的边上沿着该发送方向，将边的边向量和第一节点属性值发送给第二节点，将第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，根据最优权重的边连接的第一节点的属性值，更新第二节点的属性值；迭代执行前述步骤，直至迭代符合预设停止条件，根据迭代后的各节点的属性值，确认同在目标团的各节点，以及确认目标团中各节点的属性。

Description

团网络识别方法、装置、计算机设备及计算机可读存储介质
本申请要求于2018年04月04日提交中国专利局，申请号为2018102984431，申请名称为“团网络识别方法、装置、服务器及计算机可读存储介质”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
本申请涉及数据处理领域，尤其涉及一种团网络识别方法、装置、计算机设备及计算机可读存储介质。
背景技术
随着关系网信息的发展，不同类型的团在数据传输关系网中逐渐形成，在数据传输关系网中挖掘具有目标特性的团，对关系网信息的监管有着极为重要的意义。
目前，在数据传输关系网中进行子关系网挖掘主要通过标签扩散算法实现，这种算法普遍存在将具有目标特性的团中的数据传输和不具有目标特性的团中的数据传输混淆的问题。
发明内容
根据本申请提供的各种实施例，提供一种团网络识别方法、装置、服务器及计算机可读存储介质。
本申请实施例一方面提供了一种团网络识别方法，包括：
计算机设备根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
所述计算机设备在构建的数据传输关系网中，在第一节点与相邻的第二节 点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
所述计算机设备将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
所述计算机设备根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
所述计算机设备迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
所述计算机设备根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
本申请实施例一方面提供了一种团网络识别装置，包括：
确认模块，用于根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
发送模块，用于在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
计算模块，用于将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
更新模块，用于根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
迭代模块，用于触发迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算， 得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
所述确认模块，还用于根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
本申请实施例一方面提供了一种计算机设备，包括存储器和处理器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行上述本申请实施例提供的团网络识别方法。
本申请实施例一方面提供了一个或多个存储有计算机可读指令的非易失性存储介质，所述计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述本申请实施例提供的团网络识别方法。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征、目的和优点将从说明书、附图以及权利要求书变得明显。
附图说明
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为本申请实施例提供的团网络识别方法的应用环境图；
图2为本申请实施例提供的团网络识别方法中的数据传输关系网的示意图；
图3为本申请实施例提供的团网络识别方法的流程图；
图4为本申请实施例提供的团网络识别方法的流程图；
图5为本申请实施例提供的团网络识别方法中虚拟资源转移关系网的示意图；
图6为图5中的虚拟资源转移关系网，在分配了初始的团核心指数后的示意图；
图7为图6中的虚拟资源转移关系网，在发送多维向量时的示意图；
图8为从图7中的虚拟资源转移关系网中识别的目标团的示意图；
图9为本申请实施例提供的团网络识别装置的结构示意图；
图10为本申请实施例提供的团网络识别装置的结构示意图；
图11为本申请实施例提供的计算机设备的硬件结构示意图。
具体实施方式
为使得本申请的发明目的、特性、优点能够更加的明显和易懂，下面将结合本申请实施例提供的附图，对本申请实施例提供的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而非全部实施例。基于本申请提供的实施例，本领域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。
请参阅图1，为本申请实施例提供的团网络识别方法的应用环境图。如图1所示，多个终端100构成一个数据传输关系网，存储服务器200记录数据传输关系网中各个终端100之间进行数据交互的相关信息数据，服务器300根据存储服务器200记录的相关信息数据，通过下述各实施例提供的团网络识别方法，从该数据传输关系网中识别出目标团。在实际应用中，服务器300可以是pregel图计算框架的分布式服务器集群。服务器200也可以是分布式服务器集群。pregel图计算框架是分布式图计算的计算框架，图由节点和边组成。
其中，数据传输关系网是指由至少两个节点和各节点之间的关系共同组成的网络拓扑结构。例如，现有三个存在通讯关系的计算机设备，为了更清楚地描述这三个计算机设备之间的通讯关系，将这三个计算机设备抽象为三个节点：节点A、节点B和节点C，并用连接两个节点的边表示两个计算机设备之间的通讯关系，于是形成了一个由这三个计算机设备构成的一个数据传输关系网。存储服务器记录了相关信息数据，例如一个计算机设备与另一个计算机设备的通讯时长，在预设时间段内一个计算机设备与另一个计算机设备的通讯频率，以及一个计算机设备与另一个计算机设备进行通讯的时间段。
团(clique)是指在数据传输关系网中具有相同特性的一组节点，这组节点的节点之间形成数据转移关系。团网络是团中所有节点之间由于数据传输关系形成的网络拓扑结构。例如，在具有虚拟资源转移关系的数据传输关系网中，节点为转移虚拟资源的账户对应的计算机设备，在节点之间进行虚拟资源转移时，备注的虚拟资源转移关键词为“押大、赌资、筹码”等与赌博相关的词语，则对应地，这些节点构成一个赌博性质的团。这些节点与节点之间虚拟资源转移关系共同构成了团网络。
请参阅图2，图2为本申请实施例提供的团网络识别方法中的数据传输关系网的示意图。本申请各实施例中的数据传输关系网，具体可以是虚拟资源转移关系网，但不应以虚拟资源转移关系构成对本申请适用的数据传输关系网的类型进行任何限定。虚拟资源转移关系网中包括多个节点，节点表示虚拟资源转出方或接收方对应的计算机设备，例如资金转出账户或资金接收账户使用的计算机设备，每条边代表该边连接的两个节点之间有数据传输，也即，两个节点之间存在虚拟资源转移关系而形成边(edge)。虚拟资源可以指可用于进行商品或服务兑换的资源，比如货币、积分、金豆、礼金券、兑换券、优惠券、礼品卡以及理财基金等等，本发明实施例对此不进行具体限定。虚拟资源可以在互联网平台上进行转移。
以图2为例，图2为一个虚拟资源转移关系网，该虚拟资源转移关系网中的节点包括：节点A、节点B和节点C。各节点之间的箭头方向表示虚拟资源转移关系，即，节点B向节点A进行虚拟资源转移，节点C向节点B进行虚拟资源转移。在实际情况中，数据传输关系网中存在大量节点，需要根据要识别出的目标团的特性，从大量节点中划分出该团，目标团可以是在虚拟资源转移关系网中，属于赌博团伙的节点构成的目标团；还可以是在客运交通运输网络中，具有通勤关系的节点构成的目标团，还可以是通讯关系网络中，具有通讯业务的节点构成的目标团。以及进一步明确该团中各节点的属性，属性可以是指该节点在团中的核心程度。
下面在各实施例中详细说明团网络识别方法、装置、计算机设备和计算机可读存储介质。以下各实施例均可以利用pregel图计算框架和spark graphx中 的api进行团网络识别，在多个服务器中保存有数据传输关系网中的各节点的数据传输信息，根据该数据传输信息执行本实施中的团网络识别方法。
请参阅图3，图3为本申请实施例提供的团网络识别方法的流程图，如图3所示，该方法由服务器执行，该方法包括：
101、根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
服务器根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向。
具体地，若该待识别的目标团的数据转移特征为发散型，则服务器确认各节点的属性值发送方向为节点之间数据转移的反方向。
发散型是指针对同一个数据发送方节点存在多个数据接收方节点，例如节点A转移数据给节点B和节点C，节点B转移数据给节点D和节点E，节点C转移数据给节点F和节点G，则在由ABCDEFG组成的团中，数据从一个节点A经过节点B和节点C，转移到了四个节点D、E、F、G，该数据转移特征为发散型；上述发散型目标团的同一个数据发送方节点会对应存在至少2个数据接收方节点。
若该待识别的目标团的数据转移特征为汇聚型，则服务器确认各节点的属性值发送方向为节点之间数据转移的正方向，汇聚型是指针对同一个数据接收方节点存在多个数据发送方节点，例如节点B和节点C向节点A转移数据，节点D和节点E向节点B转移数据，节点F和节点G向节点C转移数据，则在由ABCDEFG组成的团中，数据从节点D、E、F、G四个节点经过节点B、节点C，转移到了一个节点A，数据转移特征为汇聚型；上述汇聚型目标团的同一个数据接收方节点会对应存在至少2个数据发送方节点。
在实际应用中，数据传输关系网为客运交通网络，节点为停靠站点对应的计算机设备，待识别的目标团为最具开通客运线价值的停靠站点组成的待开通站点团，则待开通站点团的交通运输数据转移的特征为汇聚型，即，客运车均向一个站点汇聚，则在识别该待开通站点团构成的网络时，确认各节点的属性值发送方向为节点之间数据转移的正方向。
102、在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边 上沿着该发送方向，将该边的边向量和该第一节点属性值发送给第二节点；
具体的，可根据图1中存储服务器200中记录的信息数据，构建数据传输关系网，构建的数据传输关系网与现实存在的数据传输网络对应。
边向量包括多个数据转移的特征值，即，该边向量用于根据至少一个单维向量描述该数据转移的特性，例如，该边向量可包括与待识别的目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移总数等单维向量。
该关键词是指与待识别的目标团的特征对应的关键词，例如，若在虚拟资源转移关系网中待识别的目标团是赌博团伙，则资源转移关键词则是与赌博相关的词语，例如：押大、赌资、筹码；在客运交通运输网络中，服务器可以以一年为统计周期，以一个月为采样分辨率，采集各站点之间一年内每个月的乘客运输关系数据，该边向量可包括一年内各站点之间乘客运输数量的均值，一年内各站点之间乘客运输数量在每个月内的分布，一年内各站点之间乘客运输数量在每个月之间的方差。
边是指因节点之间存在数据转移关系而在节点之间形成的边。例如A节点向B节点转移数据，则A节点和B节点之间形成一条边。
属性值是指表征节点属性的参数值，可以包括节点的节点ID(Identification，身份标识符)和团核心指数以及其它表征节点属性的参数值。其中，团核心指数可表征一个节点在数据传输关系网中的核心程度。
需要说明的是，在该数据传输关系网中，可以是第二节点同时向多个第一节点转移数据，或者，可以是多个第一节点同时向第二节点转移数据，所以沿着该发送方向，有多个边向量和第一节点的属性值发送给第二节点，第二节点可以是数据转移的转入方或者转出方，第二节点是属性值的接收方。发送的具体形式可以通过消息发送，在消息中包含有边向量和属性值。
由于团内数据转移关系的汇聚性和层级性，因此沿着该发送方向发送边向量，若目标团是发散型的，则用数据转移的转入方的属性去影响数据转移的转出方的属性；若目标团是汇聚型的，则用数据转移的转出方的属性去影响数据转移的转入方的属性。
因本申请实施例的算法可以是在服务器中完成，因此服务器可以将该边的边向量和第一节点属性值发送给第二节点，开始目标团的计算过程。
103、将第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，该计算逻辑与待识别的目标团的数据转移特征相匹配；
该最优权重的边是第一节点与第二节点之间的权重最大的一条边，这条边表示的数据转移关系权重最大。
计算逻辑是权重相关的计算逻辑，与待识别的目标团的数据转移特征相匹配，即，最优权重的边因计算逻辑不同而不同，而权重是与该目标团的数据转移特征相匹配的。
边向量中各数据转移的特征值的权重值是该计算逻辑赋予的，该计算逻辑所赋予数据转移的特征值的权重值是由目标团的数据转移特征匹配的。最优权重的边的确认对最终确认的团有重大影响。
具体地，该计算逻辑中预先赋予各数据转移的特征值不同的权重值，例如，数据转移频率、转移次数、转移数量分别被赋予权重值0.6、0.3和0.1，那么在第一条边上，边向量包括数据转移频率是每天1次，转移次数是4次，转移数量是100万(按10万为单位，折合为10)；在第二条边上，边向量包括数据转移频率是每天1.5次，转移次数是6次，转移数量是20万(按10万为单位，折合为2)，则分别进行加权计算，该第一条边的边权重为1*0.6+4*0.3+10*0.1＝2.8，该第二条边的边权重为1.5*0.6+6*0.3+2*0.1＝2.9，则二者相比，第二条边为最优权重的边。
104、根据该最优权重的边连接的第一节点的属性值，更新第二节点的属性值；
具体的，在该数据传输关系网中，每个第二节点有且仅有一个第一节点通过最优权重的边与该第二节点相连，根据该第一节点的属性值，更新该第二节点的属性值。
105、迭代执行步骤101至104；
106、判断迭代是否符合预设停止条件；
具体的，迭代执行在第一节点与相邻的第二节点连接的边上沿着该属性值 的发送方向，将多个该边的边向量和第一节点属性值发送给第二节点，将第二节点接收到的多个边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据最优权重的边连接的第一节点的属性值，更新第二节点的属性值的步骤，即迭代执行步骤101～104，不断更新接收边向量一方的属性值，直至迭代符合预设停止条件。
该预设停止条件可以是进行迭代的次数超过预设的迭代次数阈值，或者迭代计算的耗时超过预设的迭代时间阈值。
该预设停止条件也可以是在数据传输网络里，没有了可更新属性值的第二节点。
可以理解的，在每一轮迭代结束时，判断迭代是否符合预设停止条件，若迭代符合预设停止条件时，执行步骤107，否则，继续下一轮迭代。
107、根据迭代后的各节点的属性值，确认同在该目标团的各节点，以及确认该目标团中各节点的属性。
节点的属性值可以确认同属于一个团的节点，以此得出目标团网络的构成。
节点的属性值还可以确认在该团中各节点的属性，属性是指各节点在团中的核心程度，也即每个节点的重要程度。例如，在一个营销团网络中，根据节点的属性值可以确认该营销团中的供应商、代理商、顾客等属性。在一个赌博团网络中，根据节点的属性值可以确认该赌博团中的庄家、理赔员、参赌者等属性。
本实施例提供的团网络识别方法，通过沿着根据待识别目标团的数据转移特征确认的发送方向，将第一节点的与第二节点之间的边的边向量发送给第二节点，并对第二节点接收到边向量，根据预设计算逻辑进行加权计算，得到最优权重的边，根据最优权重的边连接的第一节点的属性值更新第二节点的属性值，多次迭代上述步骤后根据迭代后的节点的属性值确认团以及节点在图中的属性，由于该边向量包括多个描述第二节点和第一节点之间数据转移特征的特征值，因此可以更准确地分辨出具有该数据转移特征的该目标团，并且，由于迭代可以只限于用第一节点的属性值更新第二节点的属性值，这种单向计算降低了迭代的时间复杂度。
请参阅图4，图4为本申请另一实施例提供的团网络识别方法的流程图，如图所示，该方法包括：
201、初始化数据传输关系网中每个节点的属性值；
服务器初始化数据传输关系网中每个节点的属性值，属性值包括：节点ID(Identity)和团核心指数。则初始化每个节点的属性值具体为：将该第一节点的节点ID作为该第一节点的初始ID，将该第二节点的节点ID作为该第二节点的初始ID，以及，初始化第一节点和第二节点的团核心指数同为预设数值，可选地，该预设数值为0。
团核心指数可以用于衡量数据传输关系网中各节点的重要程度。具体地，团核心指数的更新规则可以预先设置，可包括在该预设数值的基础上每次增加或减少一个更新数值，该更新数值也可以自定义，例如更新数值为1，更新数值为10等，该更新数值可以是任意数值，只要在数据传输关系网全网络均设定同一个更新数值，便不影响判断一个节点的重要程度高低。
可选地，若团核心指数的更新规则为在该预设数值的基础上每次增加一个更新数值，则一个节点的团核心指数越小，该节点在团中处于越核心的位置；若团核心指数的更新方式为在该预设数值的基础上每次减少一个更新数值，则一个节点的团核心指数越大，该节点在团中处于越核心的位置。
202、根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
服务器根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向。
具体地，若该待识别的目标团的数据转移特征为发散型，则确认各节点的属性值发送方向为节点之间数据转移的反方向。
若该待识别的目标团的数据转移特征为汇聚型，则确认各节点的属性值发送方向为节点之间数据转移的正方向。
203、在数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着该发送方向，将该边的边向量和该第一节点属性值发送给第二节点；
服务器在第一节点与相邻的第二节点连接的边上沿着该发送方向，将该边的边向量和该第一节点属性值发送给第二节点。
边向量包括多个数据转移的特征值，具体可包括：数据转移的关键字与目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量。
其中，数据转移的关键字与目标团的数据转移特征中关键词匹配的次数，是指目标团的数据转移特征中存在很多关键词，而在数据转移时的附加信息中，与这些关键词匹配的次数，可以表明与该目标团性质是否相同的重要特征值，例如，目标团为营销团，其数据转移特征中关键词可包括：销售款、代理费、数额、各种货物名称，两个节点之间进行数据转移的附加信息中的关键词与这些关键词匹配上的次数越多，表示该节点越可能属于该营销团，该附加信息可以是备注、留言等。
第一节点在预设周期内向第二节点进行的数据转移的转移频率越高，第一节点和第二节点形成的数据转移关系权重越大，转移频率具体如一周内转移数据的天数。
第一节点在预设周期内向第二节点进行的数据转移的转移次数越多，表示第一节点和第二节点形成的数据转移关系权重越大，转移次数具体如一周内转移数据的次数。
第一节点在预设周期内向第二节点进行的数据转移的转移数量越大，表示第一节点和第二节点形成的数据转移关系权重越大，转移次数具体如一周内转移数据的总数，
204、将第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边；
可选的，将该第二节点接收到的多个边向量中数据转移的特征值，按照该预设计算逻辑中每个特征值各自对应的权重由大到小的顺序，将同类型特征值的值进行比较，值最大的特征值对应的边为最优权重的边。
可选地，该数据传输关系网为虚拟资源转移关系网，第二节点接收到的边向量为虚拟资源转移的特征向量，则将第二节点接收到的多个边向量中虚拟资源转移的特征值，按照预设计算逻辑中每个特征值各自对应的权重由大到小的顺序，将同类型特征值的值进行比较，值最大的特征值对应的边为最优权重的 边。
该多个数据转移的特征值包括数据转移的关键字与该目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量。
可选的，该将该第二节点接收到的边向量中数据转移的特征值，按照该预设计算逻辑中每个该特征值各自对应的权重由大到小的顺序，将同类型该特征值的值进行比较，值最大的特征值对应的边为最优权重的边包括：
将该数据转移的关键字与该目标团的数据转移特征中关键词匹配的次数，该预设周期内数据转移的转移频率、转移次数以及转移数量，按照该预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。可选地，多个虚拟资源转移的特征值包括虚拟资源转移的关键字与目标团的虚拟资源转移特征中关键词匹配的次数，预设周期内虚拟资源的转移频率、转移次数以及转移金额，则将第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，则为将虚拟资源转移的关键字与目标团的虚拟资源转移特征中关键词匹配的次数，预设周期内虚拟资源的转移频率、转移次数以及转移金额，按照预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
该最优权重的边是第一节点与第二节点之间的权重最大的一条边，这条边表示的虚拟资源转账关系权重最大。
可选的，计算逻辑是权重相关的计算逻辑，与待识别的目标团的虚拟资源转移特征相匹配，即，最优权重的边因计算逻辑不同而不同，而权重是与该目标团的虚拟资源转移特征相匹配的。
边向量中各向量的权重值是该计算逻辑赋予的，该计算逻辑所赋予各向量的权重值是由目标团的虚拟资源转移特征匹配的。最优权重的边的确认对最终确认的团有重大影响。
具体的，该预设计算逻辑与目标团的特性相匹配。根据要确认出的团的特性不同，预设计算逻辑不同。
205、根据最优权重的边连接的第一节点的属性值，更新第二节点的属性值；
具体的，将第二节点的节点ID，更新为最优权重的边连接的第一节点的节 点ID，以及，按照更新规则将第二节点的团核心指数，更新为最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值，该更新数值可以是预设的任意值，例如1，则更新规则可以是将第二节点的团核心指数，更新为最优权重的边连接的第一节点的团核心指数加1，或者，将第二节点的团核心指数，更新为最优权重的边连接的第一节点的团核心指数减1。
在更新规则是将第二节点的团核心指数，更新为最优权重的边连接的第一节点的团核心指数加1时，一个节点的团核心指数每增加1，该节点的节点ID同时被更新为其他节点的节点ID，则说明该节点的重要性没有此其他节点重要，因此一个节点的团核心指数越大，表示该节点的核心程度越低。团核心指数增加，核心程度则下降。
206、迭代执行步骤202至205；
207、判断迭代是否符合预设停止条件；
进一步地，迭代执行在第一节点与相邻的第二节点连接的边上沿着属性值发送方向，将多个第一节点的边向量发送给第二节点，将第二节点接收到的多个边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据最优权重的边连接的第一节点的属性值，更新第二节点的属性值的步骤，直至迭代符合预设停止条件，即，迭代执行步骤202～205，直至迭代符合预设停止条件。
可以理解的，在每一轮迭代结束时，判断迭代是否符合预设停止条件，若迭代符合预设停止条件时，执行步骤208，否则，继续下一轮迭代。
具体的，该预设的停止条件为，执行步骤202至步骤205的迭代次数达到预设的迭代次数，或执行步骤202至步骤205后数据传输关系网中的各节点的节点ID不发生改变。
需要说明的是，在迭代过程中，最优权重的边连接的第一节点的属性值中的节点ID与第二节点的节点ID相同，则不更新节点ID，也不更新团核心指数。
在迭代符合预设停止条件时，执行步骤208。
208、根据迭代后的各节点的属性值，确认同在该目标团的各节点，以及确认该目标团中各节点的属性；
节点的属性值可以确认同属于一个团的节点，以此得出目标团网络的构成。具体地，节点ID相同的节点为同一个团。
节点的属性值还可以确认在该团中各节点的属性，具体按照该更新规则以及各节点的团核心指数确认各节点的属性，属性是指各节点在团中的核心程度，也即每个节点的重要程度。具体的确认方式参见前述内容的相关描述，此处不再赘述。
209、根据属于同一团的节点在所述数据传输关系网中的连接关系，生成团网络图，为团网络图中不同团核心指数的节点，标注不同的预设区别特征，并输出该团网络图。
根据属于同一团的节点在该数据传输关系网中的连接关系，生成团网络图，为团网络图中不同团核心指数的节点，标注不同的预设区别特征，例如，标注不同的预设颜色，并输出该团网络图，将节点按照团核心指数可视化的进行显示，可使用户直观地查看该团的节点的核心程度。
进一步地，在输出该团网络图之前，在团网络图中每个节点上标注节点的节点ID和团核心指数，以及，在每两个连接的节点之间的边上标注边向量，方便用户在查看该团网络图时，能够明确获知该团网络图的各节点的信息以及节点之间的关系，为判断该团的性质提供必要的信息，提高判断的准确性。
为了更清楚地说明本实施例提供的团网络识别方法，下面以从一个虚拟资源转移关系网中挖掘赌博团伙关系网为例进行说明，并非对本实施例提供的团网络识别方法进行任何形式的限定。挖掘出赌博团伙后上报相关部门，对该赌博团伙进行监控可打击。
请参阅图5，图5为本申请实施例提供的团网络识别方法中虚拟资源转移关系网的示意图。如图5所示，该虚拟资源转移关系网包括13个节点：节点A、节点B、…、节点M，在以上各节点之间，基于虚拟资源转移关系产生12条边：边edge1、边edge2、…、边edge11和边edge12，每个边可传递一个边向量，每个边向量包括三个虚拟资源转移的特征值：特征值a1、特征值a2和特征值a3，这三个虚拟资源转移的特征值分别为进行虚拟资源转移的关键字和目标团的虚拟资源转移特征中关键词匹配的次数、一周内进行虚拟资源转移的次数和 一周内转移的虚拟资源的总额。若在虚拟资源转移关系网中挖掘出赌博性质的团，则特征值a1可具体为在转移虚拟资源时的备注中的关键词与赌博相关的关键词匹配上的次数，这类关键词例如可以是：“赌资”、“筹码”、“押大”等与赌博相关的词。
首先，初始化虚拟资源转移关系网中每个节点的属性值。每个节点各自的节点ID即为各自初始的节点ID，为每个节点定义一个相同的初始的团核心指数，该初始的团核心指数可以是0。其中，初始的节点ID和初始的团核心指数都标注在对应的节点旁，如图6中节点A的初始的节点ID为1，初始的团核心指数为0，则以(1，0)的形式标注在节点A旁；如图6中节点B的初始的节点ID为2，初始的团核心指数为0，则以(2，0)的形式标注在节点B旁，其他节点采用相同的形式标注。
请参阅图6，图6为图5中的虚拟资源转移关系网在分配了初始的团核心指数后的示意图，图6中每个节点的初始的节点ID均不相同，具体地，在本实例中，以每个节点的初始的团核心指数均为0，团核心指数的更新规则为每次更新均增加1为例。
若赌博团伙这一目标团的虚拟资源转移特征为发散型，则在第一节点与第二节点连接的每条边上沿着虚拟资源转移方向的反方向，第一节点以消息的形式向第二节点发送边向量和第一节点的属性值，其中，第一节点为虚拟资源转移的接收方，第二节点为虚拟资源转移的发送方，请参阅图7，图7为图6中的虚拟资源转移关系网，在发送多维向量时的示意图。
将第二节点接收到的多个消息中的边向量，按照预设计算逻辑进行加权计算，得到目标消息。
节点D接收到第一消息Msg1和第二消息Msg2。第一消息Msg1中包括：节点A的节点ID、节点A的团核心指数以及连接节点A和节点D的边的第一边向量edge1。其中，第一边向量egde1包括三个虚拟资源转移的特征值：Msg1.a1、Msg1.a2和Msg1.a3，分别表示节点D向节点A进行虚拟资源转移时备注中关键词与目标团的虚拟资源转移关键词的匹配次数，一周内转移节点D向节点A进行虚拟资源转移的次数和一周内节点D向节点A转移的虚拟资源 的总额。第二消息Msg2中包括：节点B的节点ID、节点B的团核心指数以及第二边向量edge2，第二边向量edge2包括三个虚拟资源转移的特征值：Msg2.a1、Msg2.a2和Msg2.a3，分别表示节点D向节点B进行虚拟资源转移时备注中关键词与目标团的虚拟资源转移关键词的匹配次数，一周内转移节点D向节点B进行虚拟资源转移的次数和一周内节点D向节点B转移的虚拟资源的总额。
根据赌博团伙的虚拟资源转移特性，确定对各虚拟资源转移的特征值进行加权计算的计算逻辑，权重越大的虚拟资源转移的特征值，越优先比较，反之，权重越小的虚拟资源转移的特征值，越靠后比较，在本实例中，首先比较权重最大的a1的值，即比较Msg1.a1的值和Msg2.a1的值的大小，若Msg1.a1的值大于Msg2.a1的值，且Msg1.a1的值与Msg2.a1的值之差的绝对值大于第一预设偏差值，则将发送第一消息Msg1的边作为当前最优权重边；若Msg1.a1的值小于Msg2.a1的值，且Msg1.a1的值与Msg2.a1的值之差的绝对值大于该第一预设偏差值，则将发送第二消息Msg2的边作为当前最优权重边；若Msg1.a1的值与Msg2.a1的值之差的绝对值小于该第一预设偏差值，则比较权重小于a1的a2的值，即比较Msg1.a2的值与Msg2.a2的值的大小，若Msg1.a2的值大于Msg2.a2的值，且Msg1.a2的值与Msg2.a2的值之差的绝对值大于第二预设偏差值，以及Msg1.a2的值大于预设数值时，则将发送第一消息Msg1的边作为当前最优权重边，其中，Msg1.a2的值大于预设数值表示Msg1.a2才构成有意义的虚拟资源转移的特征值；若Msg1.a2的值小于Msg2.a2的值，且Msg1.a2的值与Msg2.a2的值之差的绝对值大于第二预设偏差值，且Msg2.a2的值大于该预设数值，则将发送第二消息Msg2的边作为当前最优权重边；若Msg1.a2的值与Msg2.a2的值之差的绝对值小于该第二预设偏差值，或Msg1.a2的值和Msg2.a2的值均不大于该预设数值，则比较权重小于a1也小于a2的a3的值，即比较Msg1.a3的值与Msg2.a3的值的大小，若Msg1.a3的值大于Msg2.a3的值，且Msg1.a3的值与Msg2.a3的值之间的绝对值大于第三预设偏差值，则将发送第一消息Msg1的边作为当前最优权重边，若Msg1.a3的值小于Msg2.a3的值，且Msg1.a3的值与Msg2.a3的值之间的绝对值大于第三预设偏差值，则 将发送第二消息Msg2的边作为当前最优权重边。
在此例中经过加权计算后，得到最优权重的边为节点B与节点D之间发送第二消息Msg2的边，则将节点B的节点ID更新为节点D的节点ID，将节点B的团核心指数加1后更新为节点D的团核心指数。
请参阅图8，图8为从图7中的虚拟资源转移关系网中识别的目标团的示意图。该团网络为一个赌博团伙的虚拟资源转移关系网。根据各节点的团核心指数，可以判断这些节点在这个赌博团伙中的角色，团核心指数越小，表示该节点在该赌博团伙中的地位越重要，在本例子中，节点H、节点F、节点G和节点I的团核心指数为2，是外围的参赌者，节点E和节点D的团核心指数为1，是理赔员，节点B的团核心指数为0，是庄家。可选的，在向相关部门上报这个赌博团伙的信息时，还可以将节点B的所有虚拟资源转移行为进行汇总，上报反洗钱部门。
在本实施例中，通过沿着根据待识别目标团的数据转移特征确认的发送方向，将第一节点的与第二节点之间的边的边向量发送给第二节点，并对第二节点接收到边向量，根据预设计算逻辑进行加权计算，得到最优权重的边，根据最优权重的边连接的第一节点的属性值更新第二节点的属性值，多次迭代上述步骤后根据迭代后的节点的属性值确认团以及节点在图中的属性，由于该边向量包括多个描述第二节点和第一节点之间数据转移特征的特征值，因此可以更准确地分辨出具有该数据转移特征的该目标团，并且，由于迭代可以只限于用第一节点的属性值更新第二节点的属性值，这种单向计算降低了迭代的时间复杂度。
请参阅图9，图9为本申请一实施例提供的团网络识别装置的结构示意图，如图9所示，该团网络识别装置包括：
确认模块401，用于根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
可选的，确认模块401还用于若待识别的目标团的数据转移特征为发散型，则确认各节点的属性值发送方向为节点之间数据转移的反方向；
确认模块401还用于若待识别的目标团的数据转移特征为汇聚型，则确认 各节点的属性值发送方向为节点之间数据转移的正方向。
发送模块402，用于在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将该边的边向量和该第一节点属性值发送给该第二节点，该边向量包括多个数据转移的特征值；
计算模块403，用于将该第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，该计算逻辑与该待识别的目标团的数据转移特征相匹配；
更新模块404，用于根据该最优权重的边连接的该第一节点的属性值，更新该第二节点的属性值；
迭代模块405，用于触发迭代执行该在第一节点与相邻的第二节点连接的边上沿着该发送方向，将该边的边向量和第一节属性值发送给该第二节点，将该第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据该最优权重的边连接的该第一节点的属性值，更新该第二节点的属性值的步骤，直至迭代符合预设停止条件；
确认模块401，还用于根据迭代后的各节点的属性值，确认同在该目标团的各节点，以及确认该目标团中各节点的属性。
本实施例相关细节请参见前述图3所示实施例的描述。
在本实施例中，通过沿着根据待识别目标团的数据转移特征确认的发送方向，将第一节点的与第二节点之间的边的边向量发送给第二节点，并对第二节点接收到边向量，根据预设计算逻辑进行加权计算，得到最优权重的边，根据最优权重的边连接的第一节点的属性值更新第二节点的属性值，多次迭代上述步骤后根据迭代后的节点的属性值确认团以及节点在图中的属性，由于该边向量包括多个描述第二节点和第一节点之间数据转移特征的特征值，因此可以更准确地分辨出具有该数据转移特征的该目标团，并且，由于迭代只限于用第一节点的属性值更新第二节点的属性值，这种单向计算降低了迭代的时间复杂度。
请参阅图10，图10为本申请一实施例提供的团网络识别装置的结构示意图，与图9所示的团网络识别装置不同的是，在本实施例中：
属性值包括：节点ID和团核心指数，则该装置进一步包括：
初始化模块501用于将该第一节点的节点ID作为该第一节点的初始ID，将该第二节点的节点ID作为该第二节点的初始ID，以及，初始化该第一节点和该第二节点的团核心指数同为预设数值。
更新模块404，还用于将该第二节点的节点ID，更新为该最优权重的边连接的第一节点的节点ID，以及，按照更新规则将该第二节点的团核心指数，更新为该最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值。
确认模块401，还用于按照该更新规则以及该各节点的团核心指数，确认该各节点的属性。
计算模块403，还用于将该第二节点接收到的多个边向量中数据转移的特征值，按照该预设计算逻辑中每个该特征值各自对应的权重由大到小的顺序，将同类型该特征值的值进行比较，值最大的特征值对应的边为最优权重的边。
该多个数据转移的特征值包括数据转移的关键字与该目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移金额，则计算模块403，还用于将该数据转移的关键字与该目标团的数据转移特征中关键词匹配的次数，该预设周期内数据转移的转移频率、转移次数以及转移金额，按照该预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
进一步地，该装置还包括：生成模块502，用于根据属于同一团的节点在该数据传输关系网中的连接关系，生成团网络图；标注模块503，用于为该团网络图中不同团核心指数的节点，标注不同的预设区别特征；输出模块504，用于输出该团网络图；标注模块503，还用于在该团网络图中每个节点上标注节点的该节点ID和该团核心指数，以及，在每两个连接的节点之间的边上标注该边向量。
在本实施例中，通过沿着根据待识别目标团的数据转移特征确认的发送方向，将第一节点的与第二节点之间的边的边向量发送给第二节点，并对第二节点接收到边向量，根据预设计算逻辑进行加权计算，得到最优权重的边，根据最优权重的边连接的第一节点的属性值更新第二节点的属性值，多次迭代上述 步骤后根据迭代后的节点的属性值确认团以及节点在图中的属性，由于该边向量包括多个描述第二节点和第一节点之间数据转移特征的特征值，因此可以更准确地分辨出具有该数据转移特征的该目标团，并且，由于迭代只限于用第一节点的属性值更新第二节点的属性值，这种单向计算降低了迭代的时间复杂度。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其它实施例的相关描述。
图11示出了一个实施例中计算机设备的内部结构图。该计算机设备具体可以是图1中的服务器120。如图11所示，该计算机设备包括通过系统总线连接的处理器、存储器以及网络接口。其中，存储器包括非易失性存储介质和内存储器。该计算机设备的非易失性存储介质存储有操作系统，还可存储有计算机可读指令，该计算机可读指令被处理器执行时，可使得处理器实现团网络识别方法。该内存储器中也可储存有计算机可读指令，该计算机可读指令被处理器执行时，可使得处理器执行团网络识别方法。本领域技术人员可以理解，图11示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一个实施例中，本申请提供的团网络识别装置可以实现为一种计算机可读指令的形式，计算机可读指令可在如图11所示的计算机设备上运行。计算机设备的存储器中可存储组成该团网络识别装置的各个程序模块，比如，图9所示的确认模块401、发送模块402、计算模块403、更新模块404、迭代模块405以及编码模块912。各个程序模块构成的计算机可读指令使得处理器执行本说明书中描述的本申请各个实施例的团网络识别方法中的步骤。
进一步地，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可以设置于上述各实施例中的团网络识别装置中。该计算机可读存储介质可以是前述图11所示实施例中的存储器。该计算机可读存储介质上存储有计算机程序，该程序被处理器执行时实现前述图3至图8所示实施例中描述的团网络识别方法。进一步地，该计算机可存储介质还可以是U盘、移动硬盘、 只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，计算机可读指令可存储于一非易失性计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。所述作为分离部件说明的模块可以是或者也可以不是物理上分开的，作为模块显示的部件可以是或者也可以不是物理模块，即可以位于一个地方，或者也可以分布到多个网络模块上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
需要说明的是，对于前述的各方法实施例，为了简便描述，故将其都表述为一系列的动作组合，但是本领域技术人员应该知悉，本申请并不受所描述的动作顺序的限制，因为依据本申请，某些步骤可以采用其它顺序或者同时进行。其次，本领域技术人员也应该知悉，说明书中所描述的实施例均属于优选实施例，所涉及的动作和模块并不一定都是本申请所必须的。在各实施例中，所述“第一”、“第二”等表述仅用于命名时的区分，并非是对顺序做出任何限定。以上为对本申请所提供的团网络识别方法、装置、服务器和计算机可读存储介质的描述，对于本领域的技术人员，依据本申请实施例的思想，在具体实施方式及应用范围上均会有改变之处，综上，本说明书内容不应理解为对本申请的限制。

Claims (29)

  1. 一种团网络识别方法，包括：
    计算机设备根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
    所述计算机设备在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
    所述计算机设备将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
    所述计算机设备根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
    所述计算机设备迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
    所述计算机设备根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
  2. 如权利要求1所述的团网络识别方法，所述属性值包括：节点ID和团核心指数，则所述计算机设备在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点之前，包括：
    所述计算机设备将所述第一节点的节点ID作为所述第一节点的初始ID，将所述第二节点的节点ID作为所述第二节点的初始ID，以及，初始化所述第一节点和所述第二节点的团核心指数同为预设数值。
  3. 如权利要求2所述的团网络识别方法，所述计算机设备根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值，包括：
    所述计算机设备将所述第二节点的节点ID，更新为所述最优权重的边连接的第一节点的节点ID，以及，按照更新规则将所述第二节点的团核心指数，更新为所述最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值；
    则所述确认所述目标团中各节点的属性包括：
    所述计算机设备按照所述更新规则以及所述各节点的团核心指数，确认所述各节点的属性。
  4. 如权利要求3所述的团网络识别方法，所述将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边包括：
    所述计算机设备将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边。
  5. 如权利要求4所述的团网络识别方法，所述多个数据转移的特征值包括数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量，所述计算机设备将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边包括：
    所述计算机设备将所述数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，所述预设周期内数据转移的转移频率、转移次数以及转移数量，按照所述预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
  6. 如权利要求5所述的团网络识别方法，所述计算机设备根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性之后包括：
    所述计算机设备根据属于同一团的节点在所述数据传输关系网中的连接关 系，生成团网络图；
    所述计算机设备为所述团网络图中不同团核心指数的节点，标注不同的预设区别特征，并输出所述团网络图。
  7. 如权利要求6所述的团网络识别方法，所述生成团网络图之后还包括：
    所述计算机设备在所述团网络图中每个节点上标注节点的所述节点ID和所述团核心指数，以及，在每两个连接的节点之间的边上标注所述边向量。
  8. 根据权利要求1所述的方法，所述计算机设备根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向包括：
    若所述待识别的目标团的数据转移特征为发散型，则所述计算机设备确认各节点的属性值发送方向为节点之间数据转移的反方向；
    若所述待识别的目标团的数据转移特征为汇聚型，则所述计算机设备确认各节点的属性值发送方向为节点之间数据转移的正方向。
  9. 一种团网络识别装置，包括：
    确认模块，用于根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
    发送模块，用于在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
    计算模块，用于将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
    更新模块，用于根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
    迭代模块，用于触发迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
    所述确认模块，还用于根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
  10. 如权利要求9所述的装置，所述属性值包括：节点ID和团核心指数，则所述装置还包括：
    初始化模块，用于将所述第一节点的节点ID作为所述第一节点的初始ID，将所述第二节点的节点ID作为所述第二节点的初始ID，以及，初始化所述第一节点和所述第二节点的团核心指数同为预设数值。
  11. 如权利要求10所述的装置，所述更新模块，还用于将所述第二节点的节点ID，更新为所述最优权重的边连接的第一节点的节点ID，以及，按照更新规则将所述第二节点的团核心指数，更新为所述最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值；
    所述确认模块，还用于按照所述更新规则以及所述各节点的团核心指数，确认所述各节点的属性。
  12. 如权利要求11所述的装置，所述计算模块，还用于将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边；
    所述多个数据转移的特征值包括数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量，则所述计算模块，还用于将所述数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，所述预设周期内数据转移的转移频率、转移次数以及转移数量，按照所述预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
  13. 如权利要求12所述的装置，所述装置还包括：
    生成模块，用于根据属于同一团的节点在所述数据传输关系网中的连接关系，生成团网络图；
    标注模块，用于为所述团网络图中不同团核心指数的节点，标注不同的预设区别特征；
    输出模块，用于输出所述团网络图；
    标注模块，还用于在所述团网络图中每个节点上标注节点的所述节点ID和所述团核心指数，以及，在每两个连接的节点之间的边上标注所述边向量。
  14. 一种计算机设备，包括存储器和处理器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述处理器执行如下步骤：
    根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
    在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
    将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
    根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
    迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
    根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
  15. 如权利要求14所述的计算机设备，所述属性值包括：节点ID和团核心指数，则所述在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点之前，所述计算机可读指令还使得所述处理器执行如下步骤：
    将所述第一节点的节点ID作为所述第一节点的初始ID，将所述第二节点的节点ID作为所述第二节点的初始ID，以及，初始化所述第一节点和所述第 二节点的团核心指数同为预设数值。
  16. 如权利要求15所述的计算机设备，所述根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值，包括：
    将所述第二节点的节点ID，更新为所述最优权重的边连接的第一节点的节点ID，以及，按照更新规则将所述第二节点的团核心指数，更新为所述最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值；
    则所述确认所述目标团中各节点的属性包括：
    按照所述更新规则以及所述各节点的团核心指数，确认所述各节点的属性。
  17. 如权利要求16所述的计算机设备，所述将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边包括：
    将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边。
  18. 如权利要求17所述的计算机设备，所述多个数据转移的特征值包括数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量，所述将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边包括：
    将所述数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，所述预设周期内数据转移的转移频率、转移次数以及转移数量，按照所述预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
  19. 如权利要求18所述的计算机设备，所述根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性之后，所述计算机可读指令还使得所述处理器执行如下步骤：
    根据属于同一团的节点在所述数据传输关系网中的连接关系，生成团网络图；
    为所述团网络图中不同团核心指数的节点，标注不同的预设区别特征，并输出所述团网络图。
  20. 如权利要求19所述的计算机设备，所述生成团网络图之后，所述计算机可读指令还使得所述处理器执行如下步骤：
    在所述团网络图中每个节点上标注节点的所述节点ID和所述团核心指数，以及，在每两个连接的节点之间的边上标注所述边向量。
  21. 根据权利要求14所述的计算机设备，所述根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向包括：
    若所述待识别的目标团的数据转移特征为发散型，则确认各节点的属性值发送方向为节点之间数据转移的反方向；
    若所述待识别的目标团的数据转移特征为汇聚型，则确认各节点的属性值发送方向为节点之间数据转移的正方向。
  22. 一个或多个存储有计算机可读指令的非易失性存储介质，所述计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行如下步骤：
    根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向；
    在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，所述边向量包括多个数据转移的特征值；
    将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边，所述计算逻辑与所述待识别的目标团的数据转移特征相匹配；
    根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值；
    迭代执行所述在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点，将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到的最优权重的边，以及根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值的步骤，直至迭代符合预设停止条件；
    根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性。
  23. 如权利要求22所述的非易失性存储介质，所述属性值包括：节点ID和团核心指数，则所述在构建的数据传输关系网中，在第一节点与相邻的第二节点连接的边上沿着所述发送方向，将所述边的边向量和所述第一节点属性值发送给所述第二节点之前，所述计算机可读指令还使得所述处理器执行如下步骤：
    将所述第一节点的节点ID作为所述第一节点的初始ID，将所述第二节点的节点ID作为所述第二节点的初始ID，以及，初始化所述第一节点和所述第二节点的团核心指数同为预设数值。
  24. 如权利要求23所述的非易失性存储介质，所述根据所述最优权重的边连接的所述第一节点的属性值，更新所述第二节点的属性值，包括：
    将所述第二节点的节点ID，更新为所述最优权重的边连接的第一节点的节点ID，以及，按照更新规则将所述第二节点的团核心指数，更新为所述最优权重的边连接的第一节点的团核心指数增加或减少预设的更新数值；
    则所述确认所述目标团中各节点的属性包括：
    按照所述更新规则以及所述各节点的团核心指数，确认所述各节点的属性。
  25. 如权利要求24所述的非易失性存储介质，所述将所述第二节点接收到的边向量，按照预设计算逻辑进行加权计算，得到最优权重的边包括：
    将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边。
  26. 如权利要求25所述的非易失性存储介质，所述多个数据转移的特征值包括数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，预设周期内数据的转移频率、转移次数以及转移数量，所述将所述第二节点接收到的多个边向量中数据转移的特征值，按照所述预设计算逻辑中每个所述特征值各自对应的权重由大到小的顺序，将同类型所述特征值的值进行比较，值最大的特征值对应的边为最优权重的边包括：
    将所述数据转移的关键字与所述目标团的数据转移特征中关键词匹配的次数，所述预设周期内数据转移的转移频率、转移次数以及转移数量，按照所述预设计算逻辑中各自对应的权重由大到小的顺序进行值的比较，值最大的特征值对应的边为最优权重的边。
  27. 如权利要求26所述的非易失性存储介质，所述根据迭代后的各节点的属性值，确认同在所述目标团的各节点，以及确认所述目标团中各节点的属性之后，所述计算机可读指令还使得所述处理器执行如下步骤：
    根据属于同一团的节点在所述数据传输关系网中的连接关系，生成团网络图；
    为所述团网络图中不同团核心指数的节点，标注不同的预设区别特征，并输出所述团网络图。
  28. 如权利要求27所述的非易失性存储介质，所述生成团网络图之后，所述计算机可读指令还使得所述处理器执行如下步骤：
    在所述团网络图中每个节点上标注节点的所述节点ID和所述团核心指数，以及，在每两个连接的节点之间的边上标注所述边向量。
  29. 根据权利要求22所述的非易失性存储介质，所述根据待识别的目标团的数据转移特征，确认各节点的属性值发送方向包括：
    若所述待识别的目标团的数据转移特征为发散型，则确认各节点的属性值发送方向为节点之间数据转移的反方向；
    若所述待识别的目标团的数据转移特征为汇聚型，则确认各节点的属性值发送方向为节点之间数据转移的正方向。
PCT/CN2019/078349 2018-04-04 2019-03-15 团网络识别方法、装置、计算机设备及计算机可读存储介质 WO2019192310A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2020545337A JP7073514B2 (ja) 2018-04-04 2019-03-15 グループネットワークの認識方法、装置、コンピュータ機器及びコンピュータ読み取り可能な媒体
US16/903,480 US10958529B2 (en) 2018-04-04 2020-06-17 Clique network identification method and apparatus, computer device, and computer-readable storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810298443.1 2018-04-04
CN201810298443.1A CN110213070B (zh) 2018-04-04 2018-04-04 团网络识别方法、装置、服务器及计算机可读存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/903,480 Continuation US10958529B2 (en) 2018-04-04 2020-06-17 Clique network identification method and apparatus, computer device, and computer-readable storage medium

Publications (1)

Publication Number Publication Date
WO2019192310A1 true WO2019192310A1 (zh) 2019-10-10

Family

ID=67778734

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/078349 WO2019192310A1 (zh) 2018-04-04 2019-03-15 团网络识别方法、装置、计算机设备及计算机可读存储介质

Country Status (4)

Country Link
US (1) US10958529B2 (zh)
JP (1) JP7073514B2 (zh)
CN (1) CN110213070B (zh)
WO (1) WO2019192310A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115134156A (zh) * 2022-06-29 2022-09-30 中国电信股份有限公司 安全等级确定方法、装置、电子设备和可读存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11971962B2 (en) * 2020-04-28 2024-04-30 Cisco Technology, Inc. Learning and assessing device classification rules
JP7448022B2 (ja) * 2020-09-16 2024-03-12 日本電信電話株式会社 検知装置、検知方法及び検知プログラム
CN112989134B (zh) * 2021-03-29 2023-08-25 腾讯科技（深圳）有限公司 节点关系图的处理方法、装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080025339A1 (en) * 2006-07-26 2008-01-31 Nokia Corporation Terminal-based contention free low overhead access
CN102346766A (zh) * 2011-09-20 2012-02-08 北京邮电大学 基于极大团发现的网络热点话题检测方法及装置
CN103368910A (zh) * 2012-03-31 2013-10-23 国际商业机器公司 虚拟无线通信网络系统及其建立方法
CN103914493A (zh) * 2013-01-09 2014-07-09 北大方正集团有限公司 一种微博用户群体结构发现分析方法及系统
CN105205184A (zh) * 2015-10-30 2015-12-30 云南大学 基于隐变量模型的用户偏好提取方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001298453A (ja) * 2000-04-14 2001-10-26 Fuji Xerox Co Ltd ネットワーク表示装置
US20100241580A1 (en) * 2009-03-19 2010-09-23 Tagged, Inc. System and method of selecting a relevant user for introduction to a user in an online environment
CN103345531B (zh) * 2013-07-26 2017-05-24 苏州大学 一种复杂网络中网络社团的确定方法及装置
CN105677648B (zh) * 2014-11-18 2018-08-28 四三九九网络股份有限公司 一种基于标签传播算法的社团发现方法及系统
CN104484344B (zh) * 2014-11-27 2017-07-28 北京理工大学 一种社会网络中社团成员层次结构的探测方法
US10129097B2 (en) * 2015-06-02 2018-11-13 ALTR Solutions, Inc. GUI and high-level API wrapper for software defined networking and software defined access for controlling network routing and rules
US10084642B2 (en) * 2015-06-02 2018-09-25 ALTR Solutions, Inc. Automated sensing of network conditions for dynamically provisioning efficient VPN tunnels
US10095878B2 (en) * 2015-06-02 2018-10-09 ALTR Solutions, Inc. Internal controls engine and reporting of events generated by a network or associated applications
CN106126521B (zh) * 2016-06-06 2018-06-19 腾讯科技（深圳）有限公司 目标对象的社交账号挖掘方法及服务器

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080025339A1 (en) * 2006-07-26 2008-01-31 Nokia Corporation Terminal-based contention free low overhead access
CN102346766A (zh) * 2011-09-20 2012-02-08 北京邮电大学 基于极大团发现的网络热点话题检测方法及装置
CN103368910A (zh) * 2012-03-31 2013-10-23 国际商业机器公司 虚拟无线通信网络系统及其建立方法
CN103914493A (zh) * 2013-01-09 2014-07-09 北大方正集团有限公司 一种微博用户群体结构发现分析方法及系统
CN105205184A (zh) * 2015-10-30 2015-12-30 云南大学 基于隐变量模型的用户偏好提取方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115134156A (zh) * 2022-06-29 2022-09-30 中国电信股份有限公司 安全等级确定方法、装置、电子设备和可读存储介质
CN115134156B (zh) * 2022-06-29 2024-01-30 中国电信股份有限公司 安全等级确定方法、装置、电子设备和可读存储介质

Also Published As

Publication number Publication date
JP2021515487A (ja) 2021-06-17
CN110213070A (zh) 2019-09-06
CN110213070B (zh) 2021-08-17
JP7073514B2 (ja) 2022-05-23
US10958529B2 (en) 2021-03-23
US20200313975A1 (en) 2020-10-01

Similar Documents

Publication Publication Date Title
WO2019192310A1 (zh) 团网络识别方法、装置、计算机设备及计算机可读存储介质
US11853724B2 (en) Graph outcome determination in domain-specific execution environment
AU2021200434B2 (en) Optimizing Neural Networks For Risk Assessment
US11210144B2 (en) Systems and methods for hyperparameter tuning
US8370280B1 (en) Combining predictive models in predictive analytical modeling
US8209274B1 (en) Predictive model importation
US20210097343A1 (en) Method and apparatus for managing artificial intelligence systems
CN112989059A (zh) 潜在客户识别方法及装置、设备及可读计算机存储介质
US9858526B2 (en) Method and system using association rules to form custom lists of cookies
CN112241805A (zh) 使用历史检验数据进行缺陷预测
US20240126730A1 (en) Schema Validation with Data Synthesis
WO2023035526A1 (zh) 对象排序方法、相关设备及介质
US11941065B1 (en) Single identifier platform for storing entity data
CN107767278B (zh) 社群层次结构构建方法和装置
Beltran-Royo et al. An effective heuristic for multistage linear programming with a stochastic right-hand side
AU2022208873B2 (en) Information matching using subgraphs
EP4293534A1 (en) Blockchain address classification method and apparatus
US10719490B1 (en) Forensic analysis using synthetic datasets
US20230297399A1 (en) Methods and systems for generating recommendations based on explainable decision trees for users of a software application
US20230171202A1 (en) System and method for performing assessment of electronic resources stored in a distributed network
Xu et al. Dissimilarity-based hypothesis testing for community detection in heterogeneous networks
CN118037284A (zh) 移动支付方法和系统
CN117745449A (zh) 风险保单识别方法、装置、计算机设备和存储介质
CN117993718A (zh) 一种基于区块链的企业风险传播路径预测方法及系统
GAO et al. Analyze Requirements for PAIS with Best Practice

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19781720

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020545337

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19781720

Country of ref document: EP

Kind code of ref document: A1