WO2021189729A1 - 复杂关系网络的信息分析方法、装置、设备及存储介质 - Google Patents
复杂关系网络的信息分析方法、装置、设备及存储介质 Download PDFInfo
- Publication number
- WO2021189729A1 WO2021189729A1 PCT/CN2020/103199 CN2020103199W WO2021189729A1 WO 2021189729 A1 WO2021189729 A1 WO 2021189729A1 CN 2020103199 W CN2020103199 W CN 2020103199W WO 2021189729 A1 WO2021189729 A1 WO 2021189729A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- community
- information
- risk value
- risk
- value
- Prior art date
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 62
- 239000013598 vector Substances 0.000 claims abstract description 103
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 75
- 238000012545 processing Methods 0.000 claims abstract description 34
- 238000006243 chemical reaction Methods 0.000 claims abstract description 19
- 238000000034 method Methods 0.000 claims abstract description 16
- 230000009467 reduction Effects 0.000 claims abstract description 13
- 238000004364 calculation method Methods 0.000 claims description 76
- 239000011159 matrix material Substances 0.000 claims description 25
- 238000012502 risk assessment Methods 0.000 claims description 10
- 238000007621 cluster analysis Methods 0.000 claims description 7
- 238000007635 classification algorithm Methods 0.000 claims description 5
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 238000005295 random walk Methods 0.000 claims description 5
- 238000009825 accumulation Methods 0.000 claims description 2
- 239000003550 marker Substances 0.000 claims 9
- 238000004590 computer program Methods 0.000 claims 6
- 230000002708 enhancing effect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9024—Graphs; Linked lists
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Definitions
- This application relates to the field of big data technology, in particular to information analysis methods, devices, equipment, and storage media for complex relational networks.
- Community discovery is an efficient method for group recognition in a complex relationship network. Many efficient community discovery algorithms have been proposed and used. In the field of anti-fraud, community discovery algorithms are also widely used, especially in the direction of fraud group identification.
- the complex relationship network is divided into communities through the community discovery algorithm, and risk propagation and risk analysis are performed on the communities obtained from the community division, and the risk value of the community is obtained.
- the inventor realizes that because the spread of risks in a complex relationship network is often limited to entities that are actually connected, only the analysis community divided by entities that are actually connected in a complex relationship network can be used for risk spreading and risk analysis.
- the associated risk value analysis is carried out between the associated communities that are not adjacent but have the same or similar characteristics as the analysis community, while for the associated communities that have the same or similar characteristics as the analysis community, it has a larger The probability of the occurrence of community risk events has a greater impact on the risk analysis of the analysis community, thus leading to weak identification and control of community group risks.
- the main purpose of this application is to solve the problem of weak identification and control of community group risks.
- the first aspect of the present application provides an information analysis method for a complex relationship network, including: obtaining a complex relationship network to be analyzed, and dividing the complex relationship network into a network topology structure through a preset algorithm to obtain The community and the community information corresponding to each of the communities; perform linear dimensionality reduction processing on the community information to obtain community features, and perform word vector conversion processing on the community features to obtain community feature vectors; calculate one of every two community feature vectors According to the target similarity, the candidate community information is determined according to the target similarity; the risk value is marked on the candidate community information to obtain the target community information, and the target community information and the target similarity are filled into the preset fields.
- Connection table generating a community weighted graph according to the connection table, the connection table is used to indicate the pointer array corresponding to the data field of the community weighting graph; obtaining the community to be evaluated and the weighted risk in the community weighting graph Value and the risk value of the community to be assessed, the weighted risk value is analyzed by a preset tag propagation algorithm to update the risk value of the community to be assessed, and the risk level and the risk value of the community to be assessed are obtained through the updated risk value.
- the weighted risk value is used to indicate a value obtained by multiplying the risk value of a community connected to or adjacent to the community to be assessed by the target similarity.
- the second aspect of the present application provides an information analysis device for a complex relational network, including a memory, a processor, and computer-readable instructions stored on the memory and running on the processor, and the processor executes all
- the computer-readable instructions implement the following steps: obtain the complex relationship network to be analyzed, divide the complex relationship network into the network topology structure through a preset algorithm, and obtain the community and the community information corresponding to each of the communities;
- the community information is subjected to linear dimensionality reduction processing to obtain community features, and word vector conversion processing is performed on the community features to obtain community feature vectors;
- the target similarity between every two community feature vectors is calculated, and the candidate is determined according to the target similarity Community information; mark the candidate community information with a risk value to obtain target community information, fill the target community information and the target similarity into a preset connection table, and generate a community weighted graph according to the connection table,
- the link table is used to indicate the pointer array corresponding to the data field of the community weighted graph; obtain the community
- the third aspect of the present application provides a computer-readable storage medium, which stores computer instructions.
- the computer executes the following steps: Obtain the complex to be analyzed The relationship network, the complex relationship network is divided into the network topology structure through a preset algorithm, and the community and the community information corresponding to each of the communities are obtained; the community information is linearly reduced to obtain the community characteristics, and the The community feature performs word vector conversion processing to obtain the community feature vector; calculates the target similarity between every two community feature vectors, and determines candidate community information according to the target similarity; performs risk value labeling on the candidate community information to obtain the target Community information, fill the target community information and the target similarity into a preset lead connection table, generate a community weighted graph according to the lead connection table, the lead connection table is used to indicate data of the community weighted graph
- the pointer array corresponding to the domain; obtain the community to be assessed, the weighted risk value, and the risk value of the community to be assessed in the community weighted graph, and
- the fourth aspect of the present application provides an information analysis device for a complex relationship network, including: a first acquisition module for acquiring a complex relationship network to be analyzed, and dividing the complex relationship network into a network topology structure through a preset algorithm , Obtain the community and the community information corresponding to each of the communities; a processing module, used to perform linear dimensionality reduction processing on the community information to obtain community features, and perform word vector conversion processing on the community features to obtain a community feature vector; calculation The module is used to calculate the target similarity between each two community feature vectors, and determine candidate community information according to the target similarity; the generating module is used to mark the candidate community information with a risk value to obtain the target community information, and The target community information and the target similarity are filled in a preset lead connection table, and a community weighted graph is generated according to the lead connection table, and the lead connection table is used to indicate the data field corresponding to the community weighted graph Pointer array; the second acquisition module is used to acquire the community to be assessed, the weighted
- the complex relationship network to be analyzed is divided into the network topology to obtain community and community information; the community information is linearly reduced to obtain community features, and the community feature vectors are obtained according to the community features; calculate every two Target similarity between two community feature vectors, determine candidate community information according to the target similarity; obtain target community information, generate a community weighted graph based on the target community information; obtain the community to be evaluated, weighted risk value, and to be evaluated in the community weighted graph
- the risk value of the community is analyzed by using a preset tag propagation algorithm to update the risk value of the community to be assessed, and the risk degree of the community to be assessed and the degree of risk identification are obtained through the updated risk value.
- a new community weighting graph is reconstructed on the basis of the complex relationship network, so as to associate the originally unrelated community groups with similarity, and adopt the label propagation algorithm Perform risk identification on the community weighted graph, obtain the spread risk value of unrelated communities, realize the analysis of risk spread between unrelated communities, and enhance the ability to identify and control the risks of community groups.
- FIG. 1 is a schematic diagram of an embodiment of an information analysis method for a complex relationship network in an embodiment of this application;
- FIG. 2 is a schematic diagram of another embodiment of a method for analyzing information of a complex relationship network in an embodiment of this application;
- FIG. 3 is a schematic diagram of an embodiment of an information analysis device for a complex relationship network in an embodiment of the application
- FIG. 4 is a schematic diagram of another embodiment of an information analysis device for a complex relationship network in an embodiment of this application;
- FIG. 5 is a schematic diagram of an embodiment of an information analysis device for a complex relationship network in an embodiment of this application.
- the embodiments of the present application provide an information analysis method, device, device, and storage medium for a complex relationship network, which are used to reconstruct a new community on the basis of the complex relationship network by using the community characteristics in the traditional complex relationship network as an intermediate variable Weighted graphs are used to associate unrelated community groups with similarity, and the label propagation algorithm is used to identify the risk of the community weighted graph to obtain the spread risk value of unrelated communities, and realize the risk spread between unrelated communities.
- the analysis has strengthened the ability to identify and control the risks of community groups.
- An embodiment of the information analysis method of the complex relationship network in the embodiment of the present application includes:
- the information analysis method of the complex relationship network includes:
- the complex relational network is divided into network topology structure through the community discovery algorithm, or the server can also group the complex relational network through preset division conditions and classification algorithms .
- the division conditions include, but are not limited to, the safety factor of the user's location and the number of users, and obtain the community and the community information corresponding to each community.
- the community information may include, but is not limited to, community node information and community related information.
- the community node information includes user information and user related information of the community, and the community related information includes related information between users in the community and related information between communities. .
- the above step 101 may include: obtaining business information and business requirements in the business information, obtaining the corresponding complex relationship network according to the business requirements; obtaining the nature of the community group in the complex relationship network; using the community discovery algorithm and the nature of the community group
- the complex relationship network is divided into the network topology structure to obtain the community; the information in the complex relationship network is obtained, and the information in the complex relationship network is classified by the classification algorithm to obtain the community information corresponding to each community.
- the information includes network topology structure information and/or sample information of a specified dimension learned based on an unsupervised learning algorithm.
- the server stores historical complex relationship network information in the form of a hash table, and the historical complex relationship network information establishes a corresponding relationship with business requirements.
- the server creates the hash value of the business demand, retrieves the hash table through the hash value, and obtains the complex relationship network corresponding to the hash value (that is, the business demand) from the hash table (that is, the historical complex relationship network information).
- the server reads the stored complex relational network through the hash table to improve the efficiency and accuracy of reading.
- the community discovery algorithm divides the complex relationship network according to the nature of the community group to divide the network topology structure, so that two unrelated groups can be directly or indirectly related through the form of a community, and the strong association and the Weakly associate two groups for division and combination.
- the execution subject of the present application may be an information analysis device of a complex relationship network, or may also be a terminal or a server, which is not specifically limited here.
- the embodiment of the present application takes the server as the execution subject as an example for description.
- the server uses a machine learning model that is a combination of a feature extraction model and a natural language processing word vector model to extract features from the community information to obtain feature information.
- the feature information is high-dimensional data, and the feature information is linearly projected In the low-dimensional space, the community features are obtained, and the features in the community information are retained with less data dimensions, and the word vector conversion processing is performed on the community features to obtain the community feature vector.
- the community feature vector is composed of multiple communities The corresponding feature vectors are combined together to form a multi-dimensional data.
- the community feature vector can be a network topology index vector.
- the network topology index vector includes the maximum degree vector, the average degree vector, the maximum correlation coefficient vector and the average correlation coefficient vector; the community feature vector can also be based on unsupervised learning algorithm learning To the feature vector of the specified dimension.
- the server calculates the target similarity between each two community feature vectors by calculating the cosine similarity algorithm or the similarity measurement algorithm calculating the Euclidean distance.
- the target similarity includes multiple similarities, and calculates the value of each similarity and the preset threshold.
- the community information is screened to obtain the community information corresponding to all the differences greater than or equal to 0, and the candidate community information is obtained.
- a community feature vector includes community feature vector 1, community feature vector 2, and community feature vector 3
- community B feature vector includes community feature vector 4, community feature vector 5, and community feature vector 6, respectively.
- Each community feature vector of A is calculated accordingly
- the similarity between the vector and the feature vector of each B community, the similarity 14, the similarity 25 and the similarity 36 are obtained respectively, and the difference between the similarity 14, the similarity 25 and the similarity 36 and the preset threshold are calculated, respectively, Difference 14, difference 25, and difference 36, difference 14, difference 25, and difference 36 are greater than or equal to 0, then the A community information corresponding to the feature vector of the A community and the community B information corresponding to the feature vector of the B community are Candidate community information.
- the above step 103 may include: calculating the target similarity between every two community feature vectors, generating a matrix of the target similarity between every two community feature vectors, to obtain a similarity matrix; The value of each element is compared and analyzed with a preset threshold, and a target similarity matrix whose value of each element is greater than the preset threshold is obtained, and the community information corresponding to the target similarity matrix is used as candidate community information.
- the server calculates the target similarity between each two community feature vectors by calculating the cosine similarity or the Euclidean distance similarity measurement algorithm, and obtains multiple similarity values, and combines the multiple similarities to generate an n ⁇ n-dimensional The similarity symmetric matrix M.
- the element M ij of the similarity matrix represents the similarity between the community c i and the community c j .
- the server sets a preset threshold in advance according to expert rules or machine learning algorithms.
- the threshold can be a matrix, and each element value in the similarity matrix is compared with a preset threshold; the preset threshold can also be a numerical value, and each element value in the similarity matrix is compared with the preset threshold .
- the community information is screened by analyzing whether the value of each element in the similarity matrix is greater than a preset threshold to obtain candidate community information.
- the server performs risk assessment on the candidate community information to obtain the risk value.
- the type of risk assessment is determined by business needs.
- the risk value is marked on the candidate community information to obtain the target community information.
- the community corresponding to the target community information is taken as the vertex.
- the target similarity value between the community corresponding to the target community information and other communities is used as the weight value, and the weight value is marked on the side of the connection connecting the community corresponding to the target community information and the community corresponding to other target community information,
- Store the vertices and the labeled weight value in the link table in the form of a pointer array and convert the link table into an undirected graph or a directed graph to obtain a community weighted graph with labeled weight values.
- the adjacency table is a collection of adjacency tables of all nodes of the community weighted graph (ie, the community corresponding to the target community information), and the adjacency table for each node is all its arcs (including the community corresponding to the target community information, the community and the The connection direction of other communities, and the mark weight value on the arc).
- the above step 104 may include: performing risk assessment on the candidate community information to obtain a risk value, establishing a corresponding relationship between the risk value and the candidate community information, and using the candidate community information for which the corresponding relationship is established as the target community information; Any target community information is used as the first community information, the target community information except the first community information in the target community information is used as the second community information, and the community feature vector corresponding to the first community information and the second community information are obtained.
- the target similarity between the community feature vectors according to the value of the target similarity between the community feature vector corresponding to the first community information and the community feature vector corresponding to the second community information, the value of the target similarity between the first community information and the second community information is from large to small
- the community information is sorted to obtain the sorting information; the communities corresponding to the first community information are used as nodes, the communities corresponding to the second community information are sequentially used as the connecting nodes of the nodes according to the sorting information, and the target similarity is used as the arc connecting the two communities Store the node, connection node, and mark weight value in the corresponding storage space in the preset connection table, and perform image element conversion on the connection table to obtain the community weighted graph.
- the server connects the communities corresponding to the target community information and marks the corresponding target similarity to construct a new graph (ie, community weighted graph), so that the similarity of the network topology between the unrelated communities in the complex relationship network Associate, so as to realize the risk spread between unrelated communities.
- a new graph ie, community weighted graph
- the communities are connected according to the target similarity sorted by the value from the largest to the smallest, so as to facilitate the subsequent spread of risks between the communities and improve their operational efficiency.
- the similarity is used as the weight value to facilitate the subsequent weighted evaluation of the risk value of the community, thereby ensuring the quality and accuracy of the risk value.
- the weighted risk value is used to indicate the value of the risk value of the community connected or adjacent to the community to be assessed multiplied by the similarity.
- the server selects the community to be evaluated from the community weighted graph based on the business requirements in the business information, and randomly walks the community weighted graph with the community to be evaluated as the starting node to obtain multiple sequence data, and screen multiple sequence data to obtain and
- the community to be assessed is connected or in the surrounding communities in the preset adjacent range, and the surrounding communities are spread risk through the tag propagation algorithm, that is, the weighted risk value of the surrounding communities is analyzed to update the risk value of the community to be assessed, according to the updated risk
- the source of hazard is the source, state and/or behavior of the risk event corresponding to the risk type of the updated risk value that may cause personal injury and/or health damage and/or property loss.
- the degree of identification and control of risks where the degree of recognition refers to the degree of identification of the degree of risk caused by each risk type in the community, and the degree of control refers to the degree of control, identification and control of the risk generated by each risk type in the community
- the degree can be expressed by a numerical value or by a degree word of high, medium or low. For example, when the risk level is 8, the hazard source is a harsh environment, and the hazard index is red, the recognition degree is high and the control degree is low.
- the risk propagation between the community to be assessed and the surrounding communities follows a principle: the greater the value of the similarity between the feature vectors of every two communities, the more similar the corresponding risk results between the two communities, that is, the risk of a certain community
- the value will propagate to the community that is most similar to its network topology, regardless of whether the two communities are adjacent in a complex relationship network, as long as the network topology of the two communities in the community weighted graph is similar, then it can be based on
- the similarity of the network topology is used to spread the risk to evaluate the recognition and control of the risk in the community to be evaluated.
- the corresponding risk value may be gradually eliminated or gradually increased, that is, the recognition and control of the risk of the community to be assessed will gradually weaken or follow.
- Gradually increase, the elimination or increase of the risk value depends on the degree of risk of other communities similar to the network topology of the community to be assessed.
- the above step 105 may include: obtaining the needs to be evaluated, traversing the weighted graph of the communities according to the needs to be evaluated, and obtaining the communities to be evaluated and the risk values of the communities to be evaluated that meet the needs to be evaluated;
- the starting node performs a random walk on the community weighted graph to obtain sequence data; obtains and marks the community corresponding to the node connected to the starting node in the sequence data, and obtains and marks the node corresponding to the node adjacent to the starting node in the target community
- the community of, get the surrounding community, the target community is used to indicate the community corresponding to the node that is not connected to the starting node in the sequence data; read the risk value and the label weight value of the surrounding community, and calculate the product of the risk value and the label weight value, Use the product as the weighted risk value of the surrounding community; substitute the weighted risk value into the preset calculation strategy of the preset tag propagation algorithm to calculate the updated risk value, and obtain the risk level and the risk level of the community
- the target community information is marked with a risk value
- the risk value of the community to be assessed is the risk value marked in the target community information corresponding to the community to be assessed.
- the server performs calculations on the community weighted graph by using at least one of the cumulative calculation strategy, the most value calculation strategy, and the crowded calculation strategy, as well as the preset calculation strategy of the preset weight ratio calculation strategy (ie, the preset label propagation algorithm).
- Risk dissemination the risk nature of each community (that is, the risk value) is disseminated as a label, so as to realize the risk dissemination between communities with similar characteristics (that is, the risk value of neighbor nodes of the community to be evaluated is calculated through a preset calculation strategy
- the target risk value of the community to be assessed the community that has not reached the performance period is further extracted for risk warning and identification.
- the spread of risks between the communities is identified and analyzed.
- the above-mentioned weighted risk value is substituted into the preset calculation strategy of the preset tag propagation algorithm for calculation, and the updated risk value is obtained, and the risk degree of the community to be assessed and the recognition degree of the risk are obtained through the updated risk value.
- It may include: accumulating and calculating the weighted risk values according to the accumulative calculation strategy to obtain the first risk value; or, according to the most value calculation strategy, sorting the weighted risk values in descending order of value, and sorting as The weighted risk value in the first order is used as the second risk value; or, according to the crowd-type calculation strategy, a new weight value is assigned to the weighted risk value, and the new weight values are sorted in descending order of value, and the new The weighted risk value ranked first in the order of the weight value is used as the third risk value; at least one of the first risk value, the second risk value and the third risk value is calculated according to the preset weighted ratio calculation strategy according to the preset weighted ratio Take the weighted value as the updated risk value of the community to be assessed, and obtain the risk degree of the community to be assessed and the degree of risk recognition through the updated risk value.
- a new community weighted graph is reconstructed on the basis of the complex relationship network, so as to associate the originally unrelated community groups with similarity, and use labels
- the dissemination algorithm performs risk identification on the community weighted graph, obtains the dissemination risk value of unrelated communities, realizes the analysis of the risk dissemination between unrelated communities, and enhances the ability to identify and control the risks of community groups.
- FIG. 2 another embodiment of the information analysis method of the complex relationship network in the embodiment of the present application includes:
- the weighted risk value is used to indicate the value of the risk value of the community connected or adjacent to the community to be assessed multiplied by the similarity;
- the methods from 201 to 205 can be referred to from 101 to 105, which will not be repeated here.
- the server performs cluster analysis on the target risk value through K-means clustering or other clustering algorithms to obtain a risk type that is the same or similar or related to the target risk value type (ie, early warning risk category), and creates the hash value of the early warning risk category , Retrieve the hash table of historical strategy schemes created by historical storage in the database through the hash value to obtain candidate strategy schemes with the same or similar hash value, or use the label extraction algorithm to label the historical strategy schemes stored in the history
- the information is extracted from the risk type to obtain the analysis risk type, the first similarity between the early warning risk category and the analysis risk type is calculated, the candidate strategy plan of the early warning risk category corresponding to the largest first similarity degree is obtained, the candidate risk value corresponding to the candidate strategy plan is calculated, and Calculate the difference between the candidate risk value and the target risk value, obtain the candidate strategy plan corresponding to the smallest difference, and use the candidate strategy plan corresponding to the smallest difference as the target strategy plan, and establish the early warning risk category, target strategy plan and target risk value.
- the relationship between the participants can be obtained by searching the early warning risk category and/or target risk value to obtain the corresponding target strategy plan.
- Cluster analysis of the target risk value by clustering algorithm to obtain the same or similar or related early warning risk category with the target risk value type, so that the potential and the risk factors that need attention can be displayed when the target risk value is displayed.
- By obtaining the corresponding strategic plan increase the multi-angle information of the target risk value.
- the target risk value is clustered through a clustering algorithm Analyze and obtain early warning risk categories that are the same or similar or related to the target risk value, so that the potential and need to pay attention to the risk factors can be displayed when the target risk value is displayed, and the target risk value can be increased by obtaining the corresponding strategic plan Of multi-angle information.
- the information analysis method of the complex relationship network in the embodiment of the application is described above.
- the information analysis device of the complex relationship network in the embodiment of the application is described below. Please refer to FIG. 3, the information analysis device of the complex relationship network in the embodiment of the application.
- One embodiment includes:
- the first obtaining module 301 is used to obtain the complex relationship network to be analyzed, divide the complex relationship network into the network topology structure through a preset algorithm, and obtain the community and the community information corresponding to each community;
- the processing module 302 is configured to perform linear dimensionality reduction processing on the community information to obtain community features, and perform word vector conversion processing on the community features to obtain community feature vectors;
- the calculation module 303 is used to calculate the target similarity between every two community feature vectors, and determine candidate community information according to the target similarity;
- the generating module 304 is used to mark the risk value of the candidate community information to obtain the target community information, fill the target community information and the target similarity into the preset connection table, and generate the community weighted graph according to the connection table, and the connection table is used for Indicates the pointer array corresponding to the data field of the community weighted graph;
- the second obtaining module 305 is used to obtain the community to be assessed, the weighted risk value, and the risk value of the community to be assessed in the community weighted graph, and to analyze the weighted risk value through a preset label propagation algorithm to update the risk value of the community to be assessed.
- the latter risk value obtains the degree of risk of the community to be assessed and the degree of recognition of the risk
- the weighted risk value is used to indicate the value of the risk value of the community connected or adjacent to the community to be assessed multiplied by the target similarity.
- a new community weighted graph is reconstructed on the basis of the complex relationship network, so as to associate the originally unrelated community groups with similarity, and use labels
- the dissemination algorithm performs risk identification on the community weighted graph, obtains the dissemination risk value of unrelated communities, realizes the analysis of the risk dissemination between unrelated communities, and enhances the ability to identify and control the risks of community groups.
- another embodiment of the information analysis device for a complex relationship network in the embodiment of the present application includes:
- the first obtaining module 301 is used to obtain the complex relationship network to be analyzed, divide the complex relationship network into the network topology structure through a preset algorithm, and obtain the community and the community information corresponding to each community;
- the processing module 302 is configured to perform linear dimensionality reduction processing on the community information to obtain community features, and perform word vector conversion processing on the community features to obtain community feature vectors;
- the calculation module 303 is used to calculate the target similarity between every two community feature vectors, and determine candidate community information according to the target similarity;
- the generating module 304 is used to mark the risk value of the candidate community information to obtain the target community information, fill the target community information and the target similarity into the preset connection table, and generate the community weighted graph according to the connection table, and the connection table is used for Indicates the pointer array corresponding to the data field of the community weighted graph;
- the second obtaining module 305 is used to obtain the community to be assessed, the weighted risk value, and the risk value of the community to be assessed in the community weighted graph, and to analyze the weighted risk value through a preset label propagation algorithm to update the risk value of the community to be assessed.
- the latter risk value obtains the degree of risk of the community to be assessed and the degree of recognition of the risk, and the weighted risk value is used to indicate the value of the risk value of the community connected or adjacent to the community to be assessed multiplied by the target similarity;
- the analysis module 306 is configured to perform cluster analysis on the target risk value through a preset clustering algorithm to obtain the early warning risk category, and obtain the corresponding target strategy plan from the historical strategy plan stored in history according to the early warning risk category and the target risk value, Establish the relationship between the early warning risk category, the target strategy plan and the target risk value.
- the generation module 304 may also be specifically used to: perform risk assessment on candidate community information to obtain a risk value, establish a correspondence between the risk value and candidate community information, and use the candidate community information for which the correspondence relationship is established as target community information; Any target community information in the community information is used as the first community information, and the target community information except the first community information in the target community information is used as the second community information, and the community feature vector corresponding to the first community information and the second community information are obtained.
- the target similarity between the community feature vectors corresponding to the community information; the value of the target similarity between the community feature vector corresponding to the first community information and the community feature vector corresponding to the second community information is from large to small to the first community
- the information and the second community information are sorted to obtain the sorting information; the communities corresponding to the first community information are used as nodes, the communities corresponding to the second community information are used as the connecting nodes of the nodes according to the sorting information, and the target similarity is regarded as connecting two
- the mark weight value on the arc of the community; store the node, the connection node and the mark weight value in the preset link table, and the link table is converted to the image element to obtain the community weighted graph.
- the second acquisition module 305 includes: a first acquisition unit 3051, configured to acquire the needs to be evaluated, traverse the community weighted graph according to the needs to be evaluated, and obtain the communities to be evaluated that meet the needs to be evaluated and the risk values of the communities to be evaluated;
- the second acquisition unit 3052 is used to take the community to be evaluated as the starting node and perform a random walk on the community weighted graph according to the starting node to obtain sequence data;
- the third acquisition unit 3053 is used to acquire and mark the sequence data
- the community corresponding to the node connected by the start node, and the community corresponding to the node adjacent to the start node is obtained and marked in the target community to obtain the surrounding community.
- the target community is used to indicate the node corresponding to the node not connected to the start node in the sequence data Community
- the first calculation unit 3054 is used to read the risk value and the mark weight value of the surrounding community, calculate the product of the risk value and the mark weight value, and use the product as the weighted risk value of the surrounding community;
- the second calculation unit 3055 is used to The weighted risk value is substituted into the preset calculation strategy of the preset label propagation algorithm for calculation, and the updated risk value is obtained.
- the updated risk value is used to obtain the degree of risk of the community to be assessed and the degree of recognition of the risk.
- the preset calculation strategy includes At least one of the cumulative calculation strategy, the most value calculation strategy, and the crowd-employment calculation strategy, as well as the preset weighted ratio calculation strategy.
- the second calculation unit 3055 may also be specifically configured to: accumulate and calculate the weighted risk value according to the accumulative calculation strategy to obtain the first risk value; or, according to the most value calculation strategy, according to the value from large to small Sort the weighted risk values in the order of, and use the weighted risk value ranked first as the second risk value; or, according to the crowd-type calculation strategy, assign a new weight value to the weighted risk value, in descending order of value Sort the new weight values in order, and sort the new weight values into the first weighted risk value as the third risk value; calculate the first risk value, the second risk value and the third risk value according to the preset weighted ratio calculation strategy
- the weighted value of at least one item in the risk value is the weighted value as the updated risk value of the community to be assessed, and the risk degree of the community to be assessed and the degree of recognition of the risk are obtained through the updated risk value.
- the calculation module 303 may also be specifically configured to: calculate the target similarity between every two community feature vectors, generate a matrix of the target similarity between every two community feature vectors, to obtain a similarity matrix; The value of each element in the degree matrix is compared and analyzed with a preset threshold to obtain a target similarity matrix whose value of each element is greater than the preset threshold, and the community information corresponding to the target similarity matrix is used as candidate community information.
- the first obtaining module 301 can also be specifically used to: obtain business information and business requirements in the business information, obtain the corresponding complex relationship network according to the business requirements; obtain the nature of the community group in the complex relationship network; and discover through the community Algorithms and the nature of community groups divide the network topology of the complex relationship network to obtain the community; obtain the information in the complex relationship network, and use the classification algorithm to classify the information in the complex relationship network to obtain the community information corresponding to each community ,
- the information in the complex relational network includes network topology information and/or sample information of a specified dimension learned based on an unsupervised learning algorithm.
- the target risk value is clustered through a clustering algorithm Analyze and obtain early warning risk categories that are the same or similar or related to the target risk value, so that the potential and need to pay attention to the risk factors can be displayed when the target risk value is displayed, and the target risk value can be increased by obtaining the corresponding strategic plan Of multi-angle information.
- FIG. 5 is a schematic structural diagram of an information analysis device for a complex relationship network provided by an embodiment of the present application.
- the information analysis device 500 for a complex relationship network may have relatively large differences due to different configurations or performances, and may include one or more A processor (central processing units, CPU) 510 (for example, one or more processors) and a memory 520, one or more storage media 530 (for example, one or more storage devices with a large amount of data) storing application programs 533 or data 532.
- the memory 520 and the storage medium 530 may be short-term storage or persistent storage.
- the program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of command operations in the information analysis device 500 of the complex relational network.
- the processor 510 may be configured to communicate with the storage medium 530, and execute a series of instruction operations in the storage medium 530 on the information analysis device 500 of a complex relational network.
- the information analysis device 500 of a complex relational network may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and/or one or more operating systems 531, For example, Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.
- Windows Serve Windows Serve
- Mac OS X Unix
- Linux FreeBSD
- FIG. 5 does not constitute a limitation on the information analysis device of the complex relationship network, and may include more or less components than shown in the figure, or a combination of some Some components, or different component arrangements.
- the present application also provides an information analysis device for a complex relationship network, including: a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor are interconnected by wires; the at least one processor The processor invokes the instructions in the memory, so that the intelligent path planning device executes the steps in the information analysis method of the complex relational network.
- the present application also provides a computer-readable storage medium.
- the computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium.
- the computer-readable storage medium stores computer instructions, and when the computer instructions are executed on the computer, the computer executes the following steps:
- connection table is used to indicate the pointer array corresponding to the data field of the community weighted graph
- the community to be assessed, the weighted risk value, and the risk value of the community to be assessed in the community weighted graph analyze the weighted risk value through a preset tag propagation algorithm to update the risk value of the community to be assessed, and update the risk value of the community to be assessed.
- the latter risk value obtains the degree of risk of the community to be assessed and the degree of recognition of the risk
- the weighted risk value is used to indicate the risk value of a community connected or adjacent to the community to be assessed multiplied by the target similarity After the value.
- the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
- the technical solution of this application essentially or the part that contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium.
- a computer device which may be a personal computer, a server, or a network device, etc.
- the aforementioned storage media include: U disk, mobile hard disk, read-only memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disk or optical disk and other media that can store program code .
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
一种复杂关系网络的信息分析方法、装置、设备及存储介质,用于增强对社区群组风险的识别和把控能力,其中,方法包括:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息(101);对社区信息进行线性降维处理得到社区特征,对社区特征进行词向量转换处理,得到社区特征向量(102);计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息(103);对候选社区信息进行风险值标记得到目标社区信息,将目标社区信息和目标相似度填充至预置的领接表,根据领接表生成社区加权图,领接表用于指示社区加权图的数据域所对应的指针数组(104);获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,加权风险值用于指示与待评估社区连接或相邻的社区的风险值乘以目标相似度后的值(105)。
Description
本申请要求于2020年3月27日提交中国专利局、申请号为202010226311.5、发明名称为“复杂关系网络的信息分析方法、装置、设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在申请中。
本申请涉及大数据技术领域,尤其涉及复杂关系网络的信息分析方法、装置、设备及存储介质。
社区发现是一种复杂关系网络进行群组识别的高效方法,许多高效的社区发现算法已被提出和运用。在反欺诈领域,社区发现算法的应用也是十分广泛,特别是在欺诈群组识别方向上。通过社区发现算法对复杂关系网络进行社区划分,对社区划分所得的社区进行风险传播和风险分析,得到社区的风险值。
发明人意识到,由于对复杂关系网络中的风险传播往往局限在有实际联系的实体之间,仅对复杂关系网络中有实际联系的实体所划分的分析社区进行风险传播和风险分析,无法对于在复杂关系网络中不是相邻的但是具有与分析社区相同或相似特征的关联社区之间进行关联的风险值分析,而对于具有与分析社区相同或相似特征的关联社区来说,具有较大的社区风险事件的发生概率,对分析社区的风险分析具有较大影响,因而,导致对社区群组风险的识别和把控能力弱。
发明内容
本申请的主要目的在于解决对社区群组风险的识别和把控能力弱的问题。
为实现上述目的,本申请第一方面提供了一种复杂关系网络的信息分析方法,包括:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
本申请第二方面提供了一种复杂关系网络的信息分析设备,包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令,所述处理器执行所述计算机可读指令时实现如下步骤:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘 以所述目标相似度后的值。
本申请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计算机指令,当所述计算机指令在计算机上运行时,使得计算机执行如下步骤:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
本申请第四方面提供了一种复杂关系网络的信息分析装置,包括:第一获取模块,用于获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;处理模块,用于对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算模块,用于计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;生成模块,用于对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;第二获取模块,用于获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
本申请提供的技术方案中,对待分析的复杂关系网络进行网络拓扑结构的划分,获得社区以及社区信息;对社区信息进行线性降维处理得到社区特征,根据社区特征得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息;获得目标社区信息,根据目标社区信息生成社区加权图;获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度。本申请中,通过将传统复杂关系网络中社区特征作为中间变量,在复杂关系网络的基础上重新构建新的社区加权图,以将原本没有关联的社区群体进行相似度关联,并采用标签传播算法对社区加权图进行风险识别,获得无关联社区的传播风险值,实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力。
图1为本申请实施例中复杂关系网络的信息分析方法的一个实施例示意图;
图2为本申请实施例中复杂关系网络的信息分析方法的另一个实施例示意图;
图3为本申请实施例中复杂关系网络的信息分析装置的一个实施例示意图;
图4为本申请实施例中复杂关系网络的信息分析装置的另一个实施例示意图;
图5为本申请实施例中复杂关系网络的信息分析设备的一个实施例示意图。
本申请实施例提供了一种复杂关系网络的信息分析方法、装置、设备及存储介质,用 于通过将传统复杂关系网络中社区特征作为中间变量,在复杂关系网络的基础上重新构建新的社区加权图,以将原本没有关联的社区群体进行相似度关联,并采用标签传播算法对社区加权图进行风险识别,获得无关联社区的传播风险值,实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力。
为了使本技术领域的人员更好地理解本申请方案,下面将结合本申请实施例中的附图,对本申请实施例进行描述。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象,而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换,以便这里描述的实施例能够以除了在这里图示或描述的内容以外的顺序实施。此外,术语“包括”或“具有”及其任何变形,意图在于覆盖不排他的包含,例如,包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元,而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
为便于理解,下面对本申请实施例的具体流程进行描述,请参阅图1,本申请实施例中复杂关系网络的信息分析方法的一个实施例包括:
在一实施例中,该复杂关系网络的信息分析方法包括:
101、获取待分析的复杂关系网络,通过预置算法对复杂关系网络进行网络拓扑结构的划分,获得社区以及每个社区对应的社区信息;
服务器从数据库中读取已存储的复杂关系网络后,通过社区发现算法对该复杂关系网络进行网络拓扑结构的划分,或者服务器也可以通过预设的划分条件和分类算法对该复杂关系网络进行分群,该划分条件包括但不限于用户所处地的安全系数和用户数量,获得社区以及每个社区对应的社区信息。其中,社区信息可包括但不限于社区节点信息和社区关联信息,社区节点信息包括社区的用户信息和用户关联信息,社区关联信息包括社区中的用户之间的关联信息和社区之间的关联信息。
具体地,上述步骤101可以包括:获取业务信息和业务信息中的业务需求,根据业务需求获得对应的复杂关系网络;获取复杂关系网络中的社区群组性质;通过社区发现算法和社区群组性质对复杂关系网络进行网络拓扑结构的划分,得到社区;获取复杂关系网络中的信息,通过分类算法对复杂关系网络中的信息进行社区分类,得到每个社区对应的社区信息,复杂关系网络中的信息包括网络拓扑结构信息和/或基于无监督学习算法学习到的指定维度的样本信息。其中,服务器以哈希表的形式存储历史复杂关系网络信息,该历史复杂关系网络信息与业务需求建立有对应关系。服务器创建业务需求的哈希值,通过哈希值检索哈希表,从哈希表(即历史复杂关系网络信息)中获取哈希值(即业务需求)对应的复杂关系网络。服务器通过哈希表读取存储的复杂关系网络,以提高读取的效率和准确性。通过社区发现算法根据社区群组性质对复杂关系网络进行网络拓扑结构的划分,以便于将无关联的两个群组通过社区的形式直接或者间接的关联起来,以及准确而有效地将强关联和弱关联两个群组进行划分和组合。
可以理解的是,本申请的执行主体可以为复杂关系网络的信息分析装置,还可以是终端或者服务器,具体此处不做限定。本申请实施例以服务器为执行主体为例进行说明。
102、对社区信息进行线性降维处理得到社区特征,对社区特征进行词向量转换处理,得到社区特征向量;
服务器获得社区信息后,通过由特征提取模型和自然语言处理词向量模型组合而成的机器学习模型对社区信息进行特征提取,获得特征信息,该特征信息为高维数据,对特征信息进行线性投影到低维的空间中,得到社区特征,以较少的数据维度保留住较多的社区 信息中的特征,对社区特征进行词向量转换处理,得到社区特征向量,社区特征向量为由多个社区对应的特征向量组合一起二成的多维数据。其中,社区特征向量可为网络拓扑结构指标向量,网络拓扑结构指标向量包括最大度数向量、平均度数向量、最大相关系数向量和平均相关系数的向量;社区特征向量也可为基于无监督学习算法学习到的指定维度的特征向量。
103、计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息;
服务器通过计算余弦相似度算法或计算欧氏距离的相似性度量算法计算每两个社区特征向量之间的目标相似度,目标相似度包括多个相似度,计算各个相似度的值与预设阈值的差值,对社区信息进行筛选,获取说是所有差值均大于或者等于0所对应的社区信息,得到候选社区信息。例如:甲社区特征向量包括社区特征向量1、社区特征向量2和社区特征向量3,乙社区特征向量包括社区特征向量4、社区特征向量5和社区特征向量6,对应地分别计算各甲社区特征向量与各乙社区特征向量之间的相似度,分别得到相似度14、相似度25和相似度36,计算相似度14、相似度25和相似度36分别与预设阈值的差值,分别得到差值14、差值25和差值36,差值14、差值25和差值36均大于或者等于0,则甲社区特征向量对应的甲社区信息以及乙社区特征向量对应的乙社区信息为候选社区信息。
具体地,上述步骤103可以包括:计算每两个社区特征向量之间的目标相似度,将每两个社区特征向量之间的目标相似度生成矩阵,得到相似度矩阵;对相似度矩阵中每个元素的值与预设阈值进行对比分析,获得每个元素的值均大于预设阈值的目标相似度矩阵,将目标相似矩阵对应的社区信息作为候选社区信息。服务器通过计算余弦相似度或计算欧氏距离的相似性度量算法计算每两个社区特征向量之间的目标相似度,得到多个相似度值,将多个相似度组合生成一个n×n维的相似度对称矩阵M,相似度矩阵中M的元素M
ij表示社区c
i和所述社区c
j之间的相似度,服务器根据专家规则或机器学习算法等来预先设置一个预设阈值,该预设阈值可为一个矩阵,相似度矩阵中的每个元素值对应预设阈值进行比较;该预设阈值也可为一个数值,相似度矩阵中的每个元素值均与该预设阈值进行比较。通过分析相似度矩阵中的每个元素值是否均大于预设阈值来对社区信息进行筛选,获得候选社区信息。
104、对候选社区信息进行风险值标记得到目标社区信息,将目标社区信息和目标相似度填充至预置的领接表,根据领接表生成社区加权图,领接表用于指示构造社区加权图的数据域所对应的指针数组;
服务器对候选社区信息进行风险评估,得到风险值,该风险评估的类型由业务需求确定,将该风险值标记在候选社区信息上,得到目标社区信息,将目标社区信息对应的社区作为顶点,以该目标社区信息对应的社区与其他的社区的目标相似度的值作为权重值,并以该权重值标记在连接该目标社区信息对应的社区与其他的目标社区信息对应的社区的连接边上,将该顶点和标记权重值以指针数组形式存储在领接表中,通过将领接表转换为无向图或有向图,从而得到一个标记权重值的社区加权图,其中,社区加权图的领接表是社区加权图的所有节点(即目标社区信息对应的社区)的邻接表的集合,而对每个节点的邻接表就是它的所有出弧(包括目标社区信息对应的社区、该社区与其他社区的连接指向,以及弧上的标记权重值)。
具体地,上述步骤104可以包括:对候选社区信息进行风险评估得到风险值,建立风险值与候选社区信息的对应关系,将建立对应关系的候选社区信息作为目标社区信息;将目标社区信息中的任意一个目标社区信息作为第一社区信息,将目标社区信息中除第一社区信息之外的目标社区信息作为第二社区信息,获取第一社区信息对应的社区特征向量与 第二社区信息对应的社区特征向量之间的目标相似度;按照第一社区信息对应的社区特征向量与第二社区信息对应的社区特征向量之间的目标相似度的值从大到小对第一社区信息和第二社区信息进行排序,获得排序信息;将第一社区信息对应的社区作为节点,根据排序信息依次将第二社区信息对应的社区依次作为节点的连接节点,将目标相似度作为连接两个社区的弧上的标记权重值;将节点、连接节点和标记权重值存储至预置的领接表中对应的存储空间,对领接表进行图像元素转换得到社区加权图。
服务器通过将目标社区信息对应的社区连接起来,并标记其对应的目标相似度,构建新图(即社区加权图),以使得在复杂关系网络中没有关联的社区之间进行网络拓扑结构相似度关联起来,从而实现无关联社区之间的风险传播,例如:在复杂关系网络中,甲社区和乙社区的地理位置关系上两者之间不存在关联,但是甲社区的网络拓扑结构和乙社区的网络拓扑结构相同或相似,则通过根据网络拓扑结构构建社区加权图将甲社区和乙社区关联起来。其中,通过按照值从大到小进行排序的目标相似度来连接社区,以便于后续对社区之间的风险传播以及提高其操作效率。通过将相似度作为权重值,以便于后续对于社区的风险值的加权求值,从而保证风险值的质量和准确性。
105、获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置的标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,加权风险值用于指示与待评估社区连接或相邻的社区的风险值乘以相似度后的值。
服务器通过业务信息中的业务需求从社区加权图中选中待评估社区,以待评估社区为开始节点对社区加权图进行随机游走,得到多个序列数据,对多个序列数据进行筛选,获得与待评估社区连接或在预设相邻范围的周边社区,通过标签传播算法对周边社区进行风险传播,即对周边社区的加权风险值进行分析以更新待评估社区的风险值,根据更新后的风险值获取危险源,危险源为更新后的风险值对应风险类型的风险事件可能导致人员伤害和(或)健康损害和(或)财产损失的根源、状态和/或行为,对更新后的风险值进行等级划分得到风险等级,对危险源进行危险指数评估,得到危险指数,遍历预置的风险判断表,根据风险判断表中的风险等级、危险源和危险指数获得待评估社区的风险程度、对于风险的识别度和把控度,其中,识别度为对社区各风险类型所造成的风险程度的辨识度,把控度为对社区各风险类型所生成的风险的管控力度,识别度和把控度可用数值表示,也可用高中低的程度词表示,例如:当风险等级为8级,危险源为环境恶劣状态,危险指数为红色时,识别度为高,把控度为低。待评估社区与周边社区之间的风险传播遵循一个原则:每两个社区特征向量之间的相似度的值越大的两个社区之间对应的风险结果也越相似,即某个社区的风险值会向着与它的网络拓扑结构最相似的社区进行传播,不管这两个社区在复杂关系网络中是否相邻,只要这两个社区在社区加权图中的网络拓扑结构相似,那么就可以依据网络拓扑结构的相似度进行风险的传播,以对待评估社区待评估社区对于风险的识别度和把控度进行评估。随着社区之间的网络拓扑结构的不断变化,对应的风险值可能会随之逐渐消除或者会随之逐渐增加,即待评估社区对于风险的识别度和把控度随之逐渐减弱或随之逐渐增强,风险值的消除或增加决于与待评估社区的网络拓扑结构相似的其他社区的风险程度。
具体地,上述步骤105可以包括:获取待评估需求,根据待评估需求遍历社区加权图,获得符合待评估需求的待评估社区以及待评估社区的风险值;将待评估社区作为起始节点,根据起始节点对社区加权图进行随机游走,获得序列数据;获取并标记序列数据中与起始节点连接的节点对应的社区,以及在目标社区中获取并标记与起始节点相邻的节点对应的社区,得到周边社区,目标社区用于指示序列数据中不与起始节点连接的节点对应的社区; 读取周边社区标记的风险值和标记权重值,计算风险值和标记权重值的乘积,将乘积作为周边社区的加权风险值;将加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,预设计算策略包括累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以及预设加权比值计算策略。其中,目标社区信息中标记有风险值,待评估社区的风险值即为待评估社区对应的目标社区信息中所标记的风险值。服务器通过以累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以及预设加权比值计算策略的预设计算策略(即预置标签传播算法)对社区加权图进行风险传播,将每个社团的风险性质(即风险值)作为标签传播出去,从而实现特征相似的社区之间的风险传播(即将通过预设计算策略对待评估社区的邻居节点的风险值进行计算所得的风险值作为待评估社区的目标风险值),进一步对未达到表现期的社区进行提取风险预警和识别。通过按照预设计算策略获取待评估社区的目标风险值,以快速而准确地获取与待评估社区有群组关联关系或者无群组关联关系的社区之间的风险传播值,从而无联系的两个社区之间的风险传播得到识别和分析。
具体地,上述的将加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度可以包括:根据累加型计算策略将加权风险值进行累加求和计算,得到第一风险值;或者,根据最值型计算策略按照值从大到小的顺序对加权风险值进行排序,将排序为第一顺位的加权风险值作为第二风险值;或者,根据众举型计算策略赋予加权风险值新的权重值,按照值从大到小的顺序对新的权重值进行排序,将新的权重值排序为第一顺位的加权风险值作为第三风险值;根据预设加权比值计算策略按照预设加权比值计算第一风险值、第二风险值和第三风险值中的至少一项的加权值,将加权值作为待评估社区的更新后的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度。通过以累加型计算策略、最值型计算策略和众举型计算策略,以便于适应各种业务场景下的对于无联系的两个社区之间的风险传播的识别和分析,以及保证其识别和分析的质量。通过按照预设加权比值计算目标风险值,以提高目标风险值的准确性。
本申请实施例中,通过将传统复杂关系网络中社区特征作为中间变量,在复杂关系网络的基础上重新构建新的社区加权图,以将原本没有关联的社区群体进行相似度关联,并采用标签传播算法对社区加权图进行风险识别,获得无关联社区的传播风险值,实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力。
请参阅图2,本申请实施例中复杂关系网络的信息分析方法的另一个实施例包括:
201、获取待分析的复杂关系网络,对复杂关系网络进行网络拓扑结构的划分,获得社区以及每个社区对应的社区信息;
202、对社区信息进行线性降维处理得到社区特征,对社区特征进行词向量转换处理,得到社区特征向量;
203、计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息;
204、对候选社区信息进行风险值标记得到目标社区信息,将目标社区信息和目标相似度填充至预置的领接表,根据领接表生成社区加权图,领接表用于指示构造社区加权图的数据域所对应的指针数组;
205、获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置的标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,加权风险值用于指示与待评估社区连接或相邻的社区的风险值乘以相似度后的值;
本申请实施例中,201至205的方法可参见101至105,此处不再赘述。
206、通过预置的聚类算法对目标风险值进行聚类分析,得到预警风险类别,根据预警风险类别和目标风险值从历史存储的历史策略方案中获取对应的目标策略方案,建立预警风险类别、目标策略方案和目标风险值三者之间的关联关系。
服务器通过K均值聚类或其他的聚类算法对目标风险值进行聚类分析获得与目标风险值的类型相同或类似或关联的风险类型(即预警风险类别),创建预警风险类别的哈希值,通过哈希值对数据库中由历史存储的创建的历史策略方案的哈希表进行检索,获得哈希值相同或相似的候选策略方案,或者通过标签提取算法对历史存储的历史策略方案的标签信息进行风险类型提取得到分析风险类型,计算预警风险类别与分析风险类型的第一相似度,获得第一相似度最大对应的预警风险类别的候选策略方案,计算候选策略方案对应的候选风险值以及计算候选风险值与目标风险值的差值,获取差值最小对应的候选策略方案,并将差值最小对应的候选策略方案作为目标策略方案,建立预警风险类别、目标策略方案和目标风险值三者之间的关联关系,通过对预警风险类别和/或目标风险值进行检索,便可获取对应的目标策略方案。通过聚类算法对目标风险值进行聚类分析,获得与目标风险值的类型相同或类似或关联的预警风险类别,以使得在展示目标风险值时能展示所潜在和所需注意的危险因素,通过获得对应的策略方案,增加目标风险值的多角度信息。
本申请实施例中,在实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力的基础上,通过聚类算法对目标风险值进行聚类分析,获得与目标风险值的类型相同或类似或关联的预警风险类别,以使得在展示目标风险值时能展示所潜在和所需注意的危险因素,通过获得对应的策略方案,增加目标风险值的多角度信息。
上面对本申请实施例中复杂关系网络的信息分析方法进行了描述,下面对本申请实施例中复杂关系网络的信息分析装置进行描述,请参阅图3,本申请实施例中复杂关系网络的信息分析装置一个实施例包括:
第一获取模块301,用于获取待分析的复杂关系网络,通过预置算法对复杂关系网络进行网络拓扑结构的划分,获得社区以及每个社区对应的社区信息;
处理模块302,用于对社区信息进行线性降维处理得到社区特征,对社区特征进行词向量转换处理,得到社区特征向量;
计算模块303,用于计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息;
生成模块304,用于对候选社区信息进行风险值标记得到目标社区信息,将目标社区信息和目标相似度填充至预置的领接表,根据领接表生成社区加权图,领接表用于指示社区加权图的数据域所对应的指针数组;
第二获取模块305,用于获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,加权风险值用于指示与待评估社区连接或相邻的社区的风险值乘以目标相似度后的值。
本申请实施例中,通过将传统复杂关系网络中社区特征作为中间变量,在复杂关系网络的基础上重新构建新的社区加权图,以将原本没有关联的社区群体进行相似度关联,并采用标签传播算法对社区加权图进行风险识别,获得无关联社区的传播风险值,实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力。
请参阅图4,本申请实施例中复杂关系网络的信息分析装置的另一个实施例包括:
第一获取模块301,用于获取待分析的复杂关系网络,通过预置算法对复杂关系网络 进行网络拓扑结构的划分,获得社区以及每个社区对应的社区信息;
处理模块302,用于对社区信息进行线性降维处理得到社区特征,对社区特征进行词向量转换处理,得到社区特征向量;
计算模块303,用于计算每两个社区特征向量之间的目标相似度,根据目标相似度确定候选社区信息;
生成模块304,用于对候选社区信息进行风险值标记得到目标社区信息,将目标社区信息和目标相似度填充至预置的领接表,根据领接表生成社区加权图,领接表用于指示社区加权图的数据域所对应的指针数组;
第二获取模块305,用于获取社区加权图中的待评估社区、加权风险值以及待评估社区的风险值,通过预置标签传播算法分析加权风险值以更新待评估社区的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,加权风险值用于指示与待评估社区连接或相邻的社区的风险值乘以目标相似度后的值;
分析模块306,用于通过预置的聚类算法对目标风险值进行聚类分析,得到预警风险类别,根据预警风险类别和目标风险值从历史存储的历史策略方案中获取对应的目标策略方案,建立预警风险类别、目标策略方案和目标风险值三者之间的关联关系。
可选的,生成模块304还可以具体用于:对候选社区信息进行风险评估得到风险值,建立风险值与候选社区信息的对应关系,将建立对应关系的候选社区信息作为目标社区信息;将目标社区信息中的任意一个目标社区信息作为第一社区信息,将目标社区信息中除第一社区信息之外的目标社区信息作为第二社区信息,获取第一社区信息对应的社区特征向量与第二社区信息对应的社区特征向量之间的目标相似度;按照第一社区信息对应的社区特征向量与第二社区信息对应的社区特征向量之间的目标相似度的值从大到小对第一社区信息和第二社区信息进行排序,获得排序信息;将第一社区信息对应的社区作为节点,根据排序信息依次将第二社区信息对应的社区作为节点的连接节点,将目标相似度作为连接两个社区的弧上的标记权重值;将节点、连接节点和标记权重值存储至预置的领接表中,对领接表进行图像元素转换得到社区加权图。
可选的,第二获取模块305包括:第一获取单元3051,用于获取待评估需求,根据待评估需求遍历社区加权图,获得符合待评估需求的待评估社区以及待评估社区的风险值;第二获取单元3052,用于将待评估社区作为起始节点,根据起始节点对社区加权图进行随机游走,获得序列数据;第三获取单元3053,用于获取并标记序列数据中与起始节点连接的节点对应的社区,以及在目标社区中获取并标记与起始节点相邻的节点对应的社区,得到周边社区,目标社区用于指示序列数据中不与起始节点连接的节点对应的社区;
第一计算单元3054,用于读取周边社区标记的风险值和标记权重值,计算风险值和标记权重值的乘积,将乘积作为周边社区的加权风险值;第二计算单元3055,用于将加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别度,预设计算策略包括累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以及预设加权比值计算策略。
可选的,第二计算单元3055还可以具体用于:根据累加型计算策略将加权风险值进行累加求和计算,得到第一风险值;或者,根据最值型计算策略按照值从大到小的顺序对加权风险值进行排序,将排序为第一顺位的加权风险值作为第二风险值;或者,根据众举型计算策略赋予加权风险值新的权重值,按照值从大到小的顺序对新的权重值进行排序,将新的权重值排序为第一顺位的加权风险值作为第三风险值;根据预设加权比值计算策略计算第一风险值、第二风险值和第三风险值中的至少一项的加权值,将加权值作为待评估社区的更新后的风险值,通过更新后的风险值获得待评估社区的风险程度和对于风险的识别 度。
可选的,计算模块303还可以具体用于:计算每两个社区特征向量之间的目标相似度,将每两个社区特征向量之间的目标相似度生成矩阵,得到相似度矩阵;对相似度矩阵中每个元素的值与预设阈值进行对比分析,获得每个元素的值均大于预设阈值的目标相似度矩阵,将目标相似矩阵对应的社区信息作为候选社区信息。
可选的,第一获取模块301还可以具体用于:获取业务信息和业务信息中的业务需求,根据业务需求获得对应的复杂关系网络;获取复杂关系网络中的社区群组性质;通过社区发现算法和社区群组性质对复杂关系网络进行网络拓扑结构的划分,得到社区;获取复杂关系网络中的信息,通过分类算法对复杂关系网络中的信息进行社区分类,得到每个社区对应的社区信息,复杂关系网络中的信息包括网络拓扑结构信息和/或基于无监督学习算法学习到的指定维度的样本信息。
本申请实施例中,在实现了对无关联社区之间的风险传播情况进行分析,增强了对社区群组风险的识别和把控能力的基础上,通过聚类算法对目标风险值进行聚类分析,获得与目标风险值的类型相同或类似或关联的预警风险类别,以使得在展示目标风险值时能展示所潜在和所需注意的危险因素,通过获得对应的策略方案,增加目标风险值的多角度信息。
上面图3和图4从模块化功能实体的角度对本申请实施例中的复杂关系网络的信息分析装置进行详细描述,下面从硬件处理的角度对本申请实施例中复杂关系网络的信息分析设备进行详细描述。
图5是本申请实施例提供的一种复杂关系网络的信息分析设备的结构示意图,该复杂关系网络的信息分析设备500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处理器(central processing units,CPU)510(例如,一个或一个以上处理器)和存储器520,一个或一个以上存储应用程序533或数据532的存储介质530(例如一个或一个以上海量存储设备)。其中,存储器520和存储介质530可以是短暂存储或持久存储。存储在存储介质530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复杂关系网络的信息分析设备500中的一系列指令操作。更进一步地,处理器510可以设置为与存储介质530通信,在复杂关系网络的信息分析设备500上执行存储介质530中的一系列指令操作。
复杂关系网络的信息分析设备500还可以包括一个或一个以上电源540,一个或一个以上有线或无线网络接口550,一个或一个以上输入输出接口560,和/或,一个或一个以上操作系统531,例如Windows Serve,Mac OS X,Unix,Linux,FreeBSD等等。本领域技术人员可以理解,图5示出的复杂关系网络的信息分析设备结构并不构成对复杂关系网络的信息分析设备的限定,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。
本申请还提供一种复杂关系网络的信息分析设备,包括:存储器和至少一个处理器,所述存储器中存储有指令,所述存储器和所述至少一个处理器通过线路互连;所述至少一个处理器调用所述存储器中的所述指令,以使得所述智能化路径规划设备执行上述复杂关系网络的信息分析方法中的步骤。
本申请还提供一种计算机可读存储介质,该计算机可读存储介质可以为非易失性计算机可读存储介质,也可以为易失性计算机可读存储介质。计算机可读存储介质存储有计算机指令,当所述计算机指令在计算机上运行时,使得计算机执行如下步骤:
获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;
对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;
计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;
对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;
获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统,装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(read-only memory,ROM)、随机存取存储器(random access memory,RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述,以上实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。
Claims (20)
- 一种复杂关系网络的信息分析方法,其中,包括:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
- 根据权利要求1所述的复杂关系网络的信息分析方法,其中,所述对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述相似度填充至预置的领接表,根据所述领接表生成社区加权图,包括:对所述候选社区信息进行风险评估得到风险值,建立所述风险值与所述候选社区信息的对应关系,将建立对应关系的候选社区信息作为目标社区信息;将所述目标社区信息中的任意一个目标社区信息作为第一社区信息,将所述目标社区信息中除所述第一社区信息之外的目标社区信息作为第二社区信息,获取所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度;按照所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度的值从大到小对所述第一社区信息和所述第二社区信息进行排序,获得排序信息;将所述第一社区信息对应的社区作为节点,根据所述排序信息依次将所述第二社区信息对应的社区作为所述节点的连接节点,将所述目标相似度作为连接两个社区的弧上的标记权重值;将所述节点、所述连接节点和所述标记权重值存储至预置的领接表中,对所述领接表进行图像元素转换得到社区加权图。
- 根据权利要求2所述的复杂关系网络的信息分析方法,其中,所述获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,包括:获取待评估需求,根据所述待评估需求遍历所述社区加权图,获得符合所述待评估需求的待评估社区以及所述待评估社区的风险值;将所述待评估社区作为起始节点,根据所述起始节点对所述社区加权图进行随机游走,获得序列数据;获取并标记所述序列数据中与所述起始节点连接的节点对应的社区,以及在目标社区中获取并标记与所述起始节点相邻的节点对应的社区,得到周边社区,所述目标社区用于指示所述序列数据中不与所述起始节点连接的节点对应的社区;读取所述周边社区标记的风险值和所述标记权重值,计算所述风险值和所述标记权重值的乘积,将所述乘积作为所述周边社区的加权风险值;将所述加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述预设计算策略包括累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以及预设加权比值计算策略。
- 根据权利要求3所述的复杂关系网络的信息分析方法,其中,所述将所述加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,包括:根据所述累加型计算策略将所述加权风险值进行累加求和计算,得到第一风险值;或者,根据所述最值型计算策略按照值从大到小的顺序对所述加权风险值进行排序,将排序为第一顺位的加权风险值作为第二风险值;或者,根据众举型计算策略赋予所述加权风险值新的权重值,按照值从大到小的顺序对所述新的权重值进行排序,将所述新的权重值排序为第一顺位的加权风险值作为第三风险值;根据所述预设加权比值计算策略计算所述第一风险值、所述第二风险值和第三风险值中的至少一项的加权值,将所述加权值作为所述待评估社区的更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度。
- 根据权利要求1-4中任意一项所述的复杂关系网络的信息分析方法,其中,在所述获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度之后,还包括:通过预置的聚类算法对所述目标风险值进行聚类分析,得到预警风险类别,根据所述预警风险类别和所述目标风险值从历史存储的历史策略方案中获取对应的目标策略方案,建立所述预警风险类别、所述目标策略方案和所述目标风险值三者之间的关联关系。
- 根据权利要求1所述的复杂关系网络的信息分析方法,其中,所述计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息,包括:计算每两个社区特征向量之间的目标相似度,将所述每两个社区特征向量之间的目标相似度生成矩阵,得到相似度矩阵;对所述相似度矩阵中每个元素的值与预设阈值进行对比分析,获得每个元素的值均大于预设阈值的目标相似度矩阵,将所述目标相似矩阵对应的社区信息作为候选社区信息。
- 根据权利要求1所述的复杂关系网络的信息分析方法,其中,所述获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息,包括:获取业务信息和所述业务信息中的业务需求,根据所述业务需求获得对应的复杂关系网络;获取所述复杂关系网络中的社区群组性质;通过社区发现算法和所述社区群组性质对所述复杂关系网络进行网络拓扑结构的划分,得到社区;获取所述复杂关系网络中的信息,通过分类算法对所述复杂关系网络中的信息进行社区分类,得到每个所述社区对应的社区信息,所述复杂关系网络中的信息包括网络拓扑结构信息和/或基于无监督学习算法学习到的指定维度的样本信息。
- 一种复杂关系网络的信息分析设备,包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令,所述处理器执行所述计算机可读指令时实现如下步骤:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的 划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
- 根据权利要求8所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:对所述候选社区信息进行风险评估得到风险值,建立所述风险值与所述候选社区信息的对应关系,将建立对应关系的候选社区信息作为目标社区信息;将所述目标社区信息中的任意一个目标社区信息作为第一社区信息,将所述目标社区信息中除所述第一社区信息之外的目标社区信息作为第二社区信息,获取所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度;按照所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度的值从大到小对所述第一社区信息和所述第二社区信息进行排序,获得排序信息;将所述第一社区信息对应的社区作为节点,根据所述排序信息依次将所述第二社区信息对应的社区作为所述节点的连接节点,将所述目标相似度作为连接两个社区的弧上的标记权重值;将所述节点、所述连接节点和所述标记权重值存储至预置的领接表中,对所述领接表进行图像元素转换得到社区加权图。
- 根据权利要求9所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:所述获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,包括:获取待评估需求,根据所述待评估需求遍历所述社区加权图,获得符合所述待评估需求的待评估社区以及所述待评估社区的风险值;将所述待评估社区作为起始节点,根据所述起始节点对所述社区加权图进行随机游走,获得序列数据;获取并标记所述序列数据中与所述起始节点连接的节点对应的社区,以及在目标社区中获取并标记与所述起始节点相邻的节点对应的社区,得到周边社区,所述目标社区用于指示所述序列数据中不与所述起始节点连接的节点对应的社区;读取所述周边社区标记的风险值和所述标记权重值,计算所述风险值和所述标记权重值的乘积,将所述乘积作为所述周边社区的加权风险值;将所述加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述预设计算策略包括累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以 及预设加权比值计算策略。
- 根据权利要求10所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:所述将所述加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,包括:根据所述累加型计算策略将所述加权风险值进行累加求和计算,得到第一风险值;或者,根据所述最值型计算策略按照值从大到小的顺序对所述加权风险值进行排序,将排序为第一顺位的加权风险值作为第二风险值;或者,根据众举型计算策略赋予所述加权风险值新的权重值,按照值从大到小的顺序对所述新的权重值进行排序,将所述新的权重值排序为第一顺位的加权风险值作为第三风险值;根据所述预设加权比值计算策略计算所述第一风险值、所述第二风险值和第三风险值中的至少一项的加权值,将所述加权值作为所述待评估社区的更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度。
- 根据权利要求8-10中任意一项所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:在所述获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度之后,还包括:通过预置的聚类算法对所述目标风险值进行聚类分析,得到预警风险类别,根据所述预警风险类别和所述目标风险值从历史存储的历史策略方案中获取对应的目标策略方案,建立所述预警风险类别、所述目标策略方案和所述目标风险值三者之间的关联关系。
- 根据权利要求8所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:计算每两个社区特征向量之间的目标相似度,将所述每两个社区特征向量之间的目标相似度生成矩阵,得到相似度矩阵;对所述相似度矩阵中每个元素的值与预设阈值进行对比分析,获得每个元素的值均大于预设阈值的目标相似度矩阵,将所述目标相似矩阵对应的社区信息作为候选社区信息。
- 根据权利要求8所述的复杂关系网络的信息分析设备,所述处理器执行所述计算机程序时还实现以下步骤:获取业务信息和所述业务信息中的业务需求,根据所述业务需求获得对应的复杂关系网络;获取所述复杂关系网络中的社区群组性质;通过社区发现算法和所述社区群组性质对所述复杂关系网络进行网络拓扑结构的划分,得到社区;获取所述复杂关系网络中的信息,通过分类算法对所述复杂关系网络中的信息进行社区分类,得到每个所述社区对应的社区信息,所述复杂关系网络中的信息包括网络拓扑结构信息和/或基于无监督学习算法学习到的指定维度的样本信息。
- 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计算机指令,当所述计算机指令在计算机上运行时,使得计算机执行如下步骤:获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处 理,得到社区特征向量;计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
- 根据权利要求15所述的计算机可读存储介质,当所述计算机指令在计算机上运行时,使得计算机还执行以下步骤:对所述候选社区信息进行风险评估得到风险值,建立所述风险值与所述候选社区信息的对应关系,将建立对应关系的候选社区信息作为目标社区信息;将所述目标社区信息中的任意一个目标社区信息作为第一社区信息,将所述目标社区信息中除所述第一社区信息之外的目标社区信息作为第二社区信息,获取所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度;按照所述第一社区信息对应的社区特征向量与所述第二社区信息对应的社区特征向量之间的目标相似度的值从大到小对所述第一社区信息和所述第二社区信息进行排序,获得排序信息;将所述第一社区信息对应的社区作为节点,根据所述排序信息依次将所述第二社区信息对应的社区作为所述节点的连接节点,将所述目标相似度作为连接两个社区的弧上的标记权重值;将所述节点、所述连接节点和所述标记权重值存储至预置的领接表中,对所述领接表进行图像元素转换得到社区加权图。
- 根据权利要求16所述的计算机可读存储介质,当所述计算机指令在计算机上运行时,使得计算机还执行以下步骤:获取待评估需求,根据所述待评估需求遍历所述社区加权图,获得符合所述待评估需求的待评估社区以及所述待评估社区的风险值;将所述待评估社区作为起始节点,根据所述起始节点对所述社区加权图进行随机游走,获得序列数据;获取并标记所述序列数据中与所述起始节点连接的节点对应的社区,以及在目标社区中获取并标记与所述起始节点相邻的节点对应的社区,得到周边社区,所述目标社区用于指示所述序列数据中不与所述起始节点连接的节点对应的社区;读取所述周边社区标记的风险值和所述标记权重值,计算所述风险值和所述标记权重值的乘积,将所述乘积作为所述周边社区的加权风险值;将所述加权风险值代入预置标签传播算法的预设计算策略中进行计算,得到更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述预设计算策略包括累加型计算策略、最值型计算策略和众举型计算策略中的至少一种,以及预设加权比值计算策略。
- 根据权利要求17所述的计算机可读存储介质,当所述计算机指令在计算机上运行时,使得计算机还执行以下步骤:根据所述累加型计算策略将所述加权风险值进行累加求和计算,得到第一风险值;或者,根据所述最值型计算策略按照值从大到小的顺序对所述加权风险值进行排序,将排序 为第一顺位的加权风险值作为第二风险值;或者,根据众举型计算策略赋予所述加权风险值新的权重值,按照值从大到小的顺序对所述新的权重值进行排序,将所述新的权重值排序为第一顺位的加权风险值作为第三风险值;根据所述预设加权比值计算策略计算所述第一风险值、所述第二风险值和第三风险值中的至少一项的加权值,将所述加权值作为所述待评估社区的更新后的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度。
- 根据权利要求15-18任意一项所述的计算机可读存储介质,当所述计算机指令在计算机上运行时,使得计算机还执行以下步骤:通过预置的聚类算法对所述目标风险值进行聚类分析,得到预警风险类别,根据所述预警风险类别和所述目标风险值从历史存储的历史策略方案中获取对应的目标策略方案,建立所述预警风险类别、所述目标策略方案和所述目标风险值三者之间的关联关系。
- 一种复杂关系网络的信息分析装置,其中,所述复杂关系网络的信息分析包括:第一获取模块,用于获取待分析的复杂关系网络,通过预置算法对所述复杂关系网络进行网络拓扑结构的划分,获得社区以及每个所述社区对应的社区信息;处理模块,用于对所述社区信息进行线性降维处理得到社区特征,对所述社区特征进行词向量转换处理,得到社区特征向量;计算模块,用于计算每两个社区特征向量之间的目标相似度,根据所述目标相似度确定候选社区信息;生成模块,用于对所述候选社区信息进行风险值标记得到目标社区信息,将所述目标社区信息和所述目标相似度填充至预置的领接表,根据所述领接表生成社区加权图,所述领接表用于指示所述社区加权图的数据域所对应的指针数组;第二获取模块,用于获取所述社区加权图中的待评估社区、加权风险值以及所述待评估社区的风险值,通过预置标签传播算法分析所述加权风险值以更新所述待评估社区的风险值,通过更新后的风险值获得所述待评估社区的风险程度和对于风险的识别度,所述加权风险值用于指示与所述待评估社区连接或相邻的社区的风险值乘以所述目标相似度后的值。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010226311.5 | 2020-03-27 | ||
CN202010226311.5A CN111401775A (zh) | 2020-03-27 | 2020-03-27 | 复杂关系网络的信息分析方法、装置、设备及存储介质 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021189729A1 true WO2021189729A1 (zh) | 2021-09-30 |
Family
ID=71431329
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/103199 WO2021189729A1 (zh) | 2020-03-27 | 2020-07-21 | 复杂关系网络的信息分析方法、装置、设备及存储介质 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN111401775A (zh) |
WO (1) | WO2021189729A1 (zh) |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114038144A (zh) * | 2021-10-12 | 2022-02-11 | 中国通信建设第三工程局有限公司 | 一种基于ai的社区安防监测系统及方法 |
CN114066619A (zh) * | 2021-11-25 | 2022-02-18 | 建信金融科技有限责任公司 | 担保圈风险确定方法、装置、电子设备和存储介质 |
CN114095225A (zh) * | 2021-11-15 | 2022-02-25 | 中国电信股份有限公司 | 安全风险评估方法、装置及存储介质 |
CN114171139A (zh) * | 2021-10-20 | 2022-03-11 | 中国航发四川燃气涡轮研究院 | 压气机叶片选材方法 |
CN114219379A (zh) * | 2022-02-22 | 2022-03-22 | 北京融信数联科技有限公司 | 一种适用于社区服务圈的资源配套评估方法、及系统 |
CN114444873A (zh) * | 2021-12-28 | 2022-05-06 | 支付宝(杭州)信息技术有限公司 | 一种风险识别方法、装置以及设备 |
CN114491078A (zh) * | 2022-02-16 | 2022-05-13 | 松立控股集团股份有限公司 | 基于知识图谱的社区项目人员落脚点和同行人员分析方法 |
CN114513426A (zh) * | 2022-03-02 | 2022-05-17 | 郑州轻工业大学 | 基于节点相似度和影响力的ccn社区划分方法 |
CN114625786A (zh) * | 2022-05-12 | 2022-06-14 | 杭银消费金融股份有限公司 | 一种基于风控技术的动态数据挖掘方法及系统 |
CN114637263A (zh) * | 2022-03-15 | 2022-06-17 | 中国石油大学(北京) | 一种异常工况实时监测方法、装置、设备及存储介质 |
CN114723269A (zh) * | 2022-03-31 | 2022-07-08 | 支付宝(杭州)信息技术有限公司 | 一种事件的风险防控方法、装置及设备 |
CN114757004A (zh) * | 2022-04-02 | 2022-07-15 | 宁波市规划设计研究院 | 城市轨道交通线网规划方法、系统及存储介质 |
CN114844903A (zh) * | 2022-04-14 | 2022-08-02 | 重庆迈斯拓科技有限公司 | 一种网络信息安全保护系统 |
CN115632840A (zh) * | 2022-10-08 | 2023-01-20 | 北京天融信网络安全技术有限公司 | 基于零信任的风险处理方法、装置、设备及存储介质 |
CN115689066A (zh) * | 2022-12-30 | 2023-02-03 | 湖南三湘银行股份有限公司 | 基于图数据算法的目标供应商风险预测方法及装置 |
CN115964582A (zh) * | 2022-11-03 | 2023-04-14 | 太平洋电信股份有限公司 | 一种网络安全风险评估方法及系统 |
CN116109630A (zh) * | 2023-04-10 | 2023-05-12 | 创域智能(常熟)网联科技有限公司 | 基于传感器采集和人工智能的图像分析方法及系统 |
CN116244284A (zh) * | 2022-12-30 | 2023-06-09 | 成都中轨轨道设备有限公司 | 一种基于立体内容的大数据处理方法 |
CN116362127A (zh) * | 2023-03-31 | 2023-06-30 | 西南大学 | 一种cad模型运动部件及运动属性联合解析方法及装置 |
CN116383883A (zh) * | 2023-04-14 | 2023-07-04 | 于诚优品信息科技(福州)有限责任公司 | 一种基于大数据的数据管理权限处理方法及系统 |
CN116663897A (zh) * | 2023-05-24 | 2023-08-29 | 中国矿业大学 | 一种基于贝叶斯网络的城市老旧社区安全风险评价方法 |
CN116761177A (zh) * | 2023-08-21 | 2023-09-15 | 云镝智慧科技有限公司 | 基于5g网关的数据采集方法及其相关装置 |
CN116882767A (zh) * | 2023-09-08 | 2023-10-13 | 之江实验室 | 一种基于不完善异构关系网络图的风险预测方法及装置 |
CN116894585A (zh) * | 2023-07-25 | 2023-10-17 | 浙江三网科技股份有限公司 | 一种应用于未来社区综合数据智能分析方法和系统 |
CN117273375A (zh) * | 2023-10-19 | 2023-12-22 | 国网安徽省电力有限公司铜陵供电公司 | 基于知识图谱的配网故障处置决策监管提升系统 |
CN117272386A (zh) * | 2023-10-10 | 2023-12-22 | 广州工程技术职业学院 | 互联网大数据信息安全加密方法及装置、设备、系统 |
CN117439995A (zh) * | 2023-12-19 | 2024-01-23 | 广东博思信息技术股份有限公司 | 信息发布方法、装置、设备及存储介质 |
CN115242482B (zh) * | 2022-07-18 | 2024-01-30 | 中国电信股份有限公司 | 越权访问风险检测方法、装置、电子设备及可读存储介质 |
CN117495083A (zh) * | 2023-10-24 | 2024-02-02 | 中交广州航道局有限公司 | 一种护岸边坡稳定性监测系统及方法 |
CN117498970A (zh) * | 2023-11-03 | 2024-02-02 | 速度科技股份有限公司 | 一种基于云平台的商业航天测控服务管理系统 |
CN117592790A (zh) * | 2024-01-19 | 2024-02-23 | 全民认证科技(杭州)有限公司 | 一种区域告警方法及系统 |
CN117789977A (zh) * | 2023-11-30 | 2024-03-29 | 华中科技大学同济医学院附属同济医院 | 一种新型压疮智能预警和预防一体化方法和系统 |
CN118133047A (zh) * | 2024-05-06 | 2024-06-04 | 贝有数(南京)信息技术有限公司 | 一种企业业务匹配分析系统、方法及存储介质 |
WO2024120166A1 (zh) * | 2022-12-08 | 2024-06-13 | 马上消费金融股份有限公司 | 数据处理方法、类别识别方法及计算机设备 |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111401775A (zh) * | 2020-03-27 | 2020-07-10 | 深圳壹账通智能科技有限公司 | 复杂关系网络的信息分析方法、装置、设备及存储介质 |
CN112017059A (zh) * | 2020-07-14 | 2020-12-01 | 北京淇瑀信息科技有限公司 | 一种分级优化的风险控制方法、装置和电子设备 |
CN111860655B (zh) * | 2020-07-23 | 2023-05-30 | 中国联合网络通信集团有限公司 | 用户的处理方法、装置和设备 |
CN112819056A (zh) * | 2021-01-25 | 2021-05-18 | 百果园技术(新加坡)有限公司 | 群控账号挖掘方法、装置、设备及存储介质 |
CN112925899B (zh) * | 2021-02-09 | 2023-06-23 | 重庆中科云从科技有限公司 | 排序模型建立方法、案件线索推荐方法、装置及介质 |
CN113344697A (zh) * | 2021-06-29 | 2021-09-03 | 平安普惠企业管理有限公司 | 团伙检测方法、装置、设备及介质 |
CN114169458B (zh) * | 2021-12-13 | 2024-07-23 | 平安科技(深圳)有限公司 | 一种欺诈人员识别方法及装置、存储介质、计算机设备 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292424A (zh) * | 2017-06-01 | 2017-10-24 | 四川新网银行股份有限公司 | 一种基于复杂社交网络的反欺诈和信用风险预测方法 |
CN108009915A (zh) * | 2017-12-21 | 2018-05-08 | 连连银通电子支付有限公司 | 一种欺诈用户社区的标记方法及相关装置 |
CN109784636A (zh) * | 2018-12-13 | 2019-05-21 | 中国平安财产保险股份有限公司 | 欺诈用户识别方法、装置、计算机设备及存储介质 |
CN110852584A (zh) * | 2019-10-30 | 2020-02-28 | 北京航空航天大学 | 一种基于复杂网络分析的风险识别方法 |
CN111401775A (zh) * | 2020-03-27 | 2020-07-10 | 深圳壹账通智能科技有限公司 | 复杂关系网络的信息分析方法、装置、设备及存储介质 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2956627A1 (en) * | 2016-01-29 | 2017-07-29 | Thomson Reuters Global Resources Unlimited Company | System and engine for seeded clustering of news events |
US10885131B2 (en) * | 2016-09-12 | 2021-01-05 | Ebrahim Bagheri | System and method for temporal identification of latent user communities using electronic content |
EP3477906B1 (en) * | 2017-10-26 | 2021-03-31 | Accenture Global Solutions Limited | Systems and methods for identifying and mitigating outlier network activity |
WO2019126224A1 (en) * | 2017-12-19 | 2019-06-27 | Visa International Service Association | Hyper-graph learner for natural language comprehension |
CN109859065A (zh) * | 2019-02-28 | 2019-06-07 | 桂林理工大学 | 基于谱聚类的多目标复杂网络社区发现方法 |
-
2020
- 2020-03-27 CN CN202010226311.5A patent/CN111401775A/zh active Pending
- 2020-07-21 WO PCT/CN2020/103199 patent/WO2021189729A1/zh active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292424A (zh) * | 2017-06-01 | 2017-10-24 | 四川新网银行股份有限公司 | 一种基于复杂社交网络的反欺诈和信用风险预测方法 |
CN108009915A (zh) * | 2017-12-21 | 2018-05-08 | 连连银通电子支付有限公司 | 一种欺诈用户社区的标记方法及相关装置 |
CN109784636A (zh) * | 2018-12-13 | 2019-05-21 | 中国平安财产保险股份有限公司 | 欺诈用户识别方法、装置、计算机设备及存储介质 |
CN110852584A (zh) * | 2019-10-30 | 2020-02-28 | 北京航空航天大学 | 一种基于复杂网络分析的风险识别方法 |
CN111401775A (zh) * | 2020-03-27 | 2020-07-10 | 深圳壹账通智能科技有限公司 | 复杂关系网络的信息分析方法、装置、设备及存储介质 |
Non-Patent Citations (2)
Title |
---|
LUO GANG , ZHANG YAWEI , WANG YONG: "Risk Propagation Patterns of Guarantee Network based on the Theory of Complex Network", JOURNAL OF UNIVERSITY OF CHINESE ACADEMY OF SCIENCES, vol. 32, no. 6, 15 November 2015 (2015-11-15), pages 836 - 842, XP055853366, ISSN: 2095-6134, DOI: 10.7523/j.issn.2095-6134.2015.06.016 * |
STEINHAEUSER, K. ; CHAWLA, N.V.: "Identifying and evaluating community structure in complex networks", PATTERN RECOGNITION LETTERS, vol. 31, no. 5, 1 April 2010 (2010-04-01), NL, pages 413 - 421, XP026880476, ISSN: 0167-8655, DOI: 10.1016/j.patrec.2009.11.001 * |
Cited By (53)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114038144A (zh) * | 2021-10-12 | 2022-02-11 | 中国通信建设第三工程局有限公司 | 一种基于ai的社区安防监测系统及方法 |
CN114038144B (zh) * | 2021-10-12 | 2023-04-14 | 中国通信建设第三工程局有限公司 | 一种基于ai的社区安防监测系统及方法 |
CN114171139A (zh) * | 2021-10-20 | 2022-03-11 | 中国航发四川燃气涡轮研究院 | 压气机叶片选材方法 |
CN114095225A (zh) * | 2021-11-15 | 2022-02-25 | 中国电信股份有限公司 | 安全风险评估方法、装置及存储介质 |
CN114066619A (zh) * | 2021-11-25 | 2022-02-18 | 建信金融科技有限责任公司 | 担保圈风险确定方法、装置、电子设备和存储介质 |
CN114444873A (zh) * | 2021-12-28 | 2022-05-06 | 支付宝(杭州)信息技术有限公司 | 一种风险识别方法、装置以及设备 |
CN114491078A (zh) * | 2022-02-16 | 2022-05-13 | 松立控股集团股份有限公司 | 基于知识图谱的社区项目人员落脚点和同行人员分析方法 |
CN114491078B (zh) * | 2022-02-16 | 2022-08-02 | 松立控股集团股份有限公司 | 基于知识图谱的社区项目人员落脚点和同行人员分析方法 |
CN114219379A (zh) * | 2022-02-22 | 2022-03-22 | 北京融信数联科技有限公司 | 一种适用于社区服务圈的资源配套评估方法、及系统 |
CN114513426A (zh) * | 2022-03-02 | 2022-05-17 | 郑州轻工业大学 | 基于节点相似度和影响力的ccn社区划分方法 |
CN114513426B (zh) * | 2022-03-02 | 2023-09-15 | 郑州轻工业大学 | 基于节点相似度和影响力的ccn社区划分方法 |
CN114637263A (zh) * | 2022-03-15 | 2022-06-17 | 中国石油大学(北京) | 一种异常工况实时监测方法、装置、设备及存储介质 |
CN114637263B (zh) * | 2022-03-15 | 2024-01-12 | 中国石油大学(北京) | 一种异常工况实时监测方法、装置、设备及存储介质 |
CN114723269A (zh) * | 2022-03-31 | 2022-07-08 | 支付宝(杭州)信息技术有限公司 | 一种事件的风险防控方法、装置及设备 |
CN114757004A (zh) * | 2022-04-02 | 2022-07-15 | 宁波市规划设计研究院 | 城市轨道交通线网规划方法、系统及存储介质 |
CN114844903A (zh) * | 2022-04-14 | 2022-08-02 | 重庆迈斯拓科技有限公司 | 一种网络信息安全保护系统 |
CN114625786A (zh) * | 2022-05-12 | 2022-06-14 | 杭银消费金融股份有限公司 | 一种基于风控技术的动态数据挖掘方法及系统 |
CN115242482B (zh) * | 2022-07-18 | 2024-01-30 | 中国电信股份有限公司 | 越权访问风险检测方法、装置、电子设备及可读存储介质 |
CN115632840A (zh) * | 2022-10-08 | 2023-01-20 | 北京天融信网络安全技术有限公司 | 基于零信任的风险处理方法、装置、设备及存储介质 |
CN115632840B (zh) * | 2022-10-08 | 2023-07-04 | 北京天融信网络安全技术有限公司 | 基于零信任的风险处理方法、装置、设备及存储介质 |
CN115964582B (zh) * | 2022-11-03 | 2023-09-19 | 太平洋电信股份有限公司 | 一种网络安全风险评估方法及系统 |
CN115964582A (zh) * | 2022-11-03 | 2023-04-14 | 太平洋电信股份有限公司 | 一种网络安全风险评估方法及系统 |
WO2024120166A1 (zh) * | 2022-12-08 | 2024-06-13 | 马上消费金融股份有限公司 | 数据处理方法、类别识别方法及计算机设备 |
CN115689066A (zh) * | 2022-12-30 | 2023-02-03 | 湖南三湘银行股份有限公司 | 基于图数据算法的目标供应商风险预测方法及装置 |
CN115689066B (zh) * | 2022-12-30 | 2023-04-28 | 湖南三湘银行股份有限公司 | 基于图数据算法的目标供应商风险预测方法及装置 |
CN116244284B (zh) * | 2022-12-30 | 2023-11-14 | 成都中轨轨道设备有限公司 | 一种基于立体内容的大数据处理方法 |
CN116244284A (zh) * | 2022-12-30 | 2023-06-09 | 成都中轨轨道设备有限公司 | 一种基于立体内容的大数据处理方法 |
CN116362127B (zh) * | 2023-03-31 | 2024-04-19 | 西南大学 | 一种cad模型运动部件及运动属性联合解析方法及装置 |
CN116362127A (zh) * | 2023-03-31 | 2023-06-30 | 西南大学 | 一种cad模型运动部件及运动属性联合解析方法及装置 |
CN116109630A (zh) * | 2023-04-10 | 2023-05-12 | 创域智能(常熟)网联科技有限公司 | 基于传感器采集和人工智能的图像分析方法及系统 |
CN116383883A (zh) * | 2023-04-14 | 2023-07-04 | 于诚优品信息科技(福州)有限责任公司 | 一种基于大数据的数据管理权限处理方法及系统 |
CN116383883B (zh) * | 2023-04-14 | 2023-12-22 | 水发科技信息(山东)有限公司 | 一种基于大数据的数据管理权限处理方法及系统 |
CN116663897A (zh) * | 2023-05-24 | 2023-08-29 | 中国矿业大学 | 一种基于贝叶斯网络的城市老旧社区安全风险评价方法 |
CN116663897B (zh) * | 2023-05-24 | 2023-11-14 | 中国矿业大学 | 一种基于贝叶斯网络的城市老旧社区安全风险评价方法 |
CN116894585A (zh) * | 2023-07-25 | 2023-10-17 | 浙江三网科技股份有限公司 | 一种应用于未来社区综合数据智能分析方法和系统 |
CN116894585B (zh) * | 2023-07-25 | 2024-02-20 | 浙江三网科技股份有限公司 | 一种应用于未来社区综合数据智能分析方法和系统 |
CN116761177A (zh) * | 2023-08-21 | 2023-09-15 | 云镝智慧科技有限公司 | 基于5g网关的数据采集方法及其相关装置 |
CN116761177B (zh) * | 2023-08-21 | 2023-10-20 | 云镝智慧科技有限公司 | 基于5g网关的数据采集方法及其相关装置 |
CN116882767B (zh) * | 2023-09-08 | 2024-01-05 | 之江实验室 | 一种基于不完善异构关系网络图的风险预测方法及装置 |
CN116882767A (zh) * | 2023-09-08 | 2023-10-13 | 之江实验室 | 一种基于不完善异构关系网络图的风险预测方法及装置 |
CN117272386A (zh) * | 2023-10-10 | 2023-12-22 | 广州工程技术职业学院 | 互联网大数据信息安全加密方法及装置、设备、系统 |
CN117272386B (zh) * | 2023-10-10 | 2024-02-27 | 广州工程技术职业学院 | 互联网大数据信息安全加密方法及装置、设备、系统 |
CN117273375B (zh) * | 2023-10-19 | 2024-04-02 | 国网安徽省电力有限公司铜陵供电公司 | 基于知识图谱的配网故障处置决策监管提升系统 |
CN117273375A (zh) * | 2023-10-19 | 2023-12-22 | 国网安徽省电力有限公司铜陵供电公司 | 基于知识图谱的配网故障处置决策监管提升系统 |
CN117495083A (zh) * | 2023-10-24 | 2024-02-02 | 中交广州航道局有限公司 | 一种护岸边坡稳定性监测系统及方法 |
CN117495083B (zh) * | 2023-10-24 | 2024-04-12 | 中交广州航道局有限公司 | 一种护岸边坡稳定性监测系统及方法 |
CN117498970A (zh) * | 2023-11-03 | 2024-02-02 | 速度科技股份有限公司 | 一种基于云平台的商业航天测控服务管理系统 |
CN117789977A (zh) * | 2023-11-30 | 2024-03-29 | 华中科技大学同济医学院附属同济医院 | 一种新型压疮智能预警和预防一体化方法和系统 |
CN117439995B (zh) * | 2023-12-19 | 2024-03-15 | 广东博思信息技术股份有限公司 | 信息发布方法、装置、设备及存储介质 |
CN117439995A (zh) * | 2023-12-19 | 2024-01-23 | 广东博思信息技术股份有限公司 | 信息发布方法、装置、设备及存储介质 |
CN117592790A (zh) * | 2024-01-19 | 2024-02-23 | 全民认证科技(杭州)有限公司 | 一种区域告警方法及系统 |
CN117592790B (zh) * | 2024-01-19 | 2024-04-09 | 全民认证科技(杭州)有限公司 | 一种区域告警方法及系统 |
CN118133047A (zh) * | 2024-05-06 | 2024-06-04 | 贝有数(南京)信息技术有限公司 | 一种企业业务匹配分析系统、方法及存储介质 |
Also Published As
Publication number | Publication date |
---|---|
CN111401775A (zh) | 2020-07-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021189729A1 (zh) | 复杂关系网络的信息分析方法、装置、设备及存储介质 | |
CN105469096B (zh) | 一种基于哈希二值编码的特征袋图像检索方法 | |
WO2021189730A1 (zh) | 检测异常高密子图的方法、装置、设备及存储介质 | |
Macskassy et al. | Classification in networked data: A toolkit and a univariate case study. | |
CN108363821A (zh) | 一种信息推送方法、装置、终端设备及存储介质 | |
CN104424296B (zh) | 查询词分类方法和装置 | |
US11971892B2 (en) | Methods for stratified sampling-based query execution | |
CN111553127A (zh) | 一种多标记的文本类数据特征选择方法及装置 | |
CN104239553A (zh) | 一种基于Map-Reduce框架的实体识别方法 | |
CN102799614B (zh) | 基于视觉词语空间共生性的图像检索方法 | |
CN111612041A (zh) | 异常用户识别方法及装置、存储介质、电子设备 | |
Barz et al. | Enhancing flood impact analysis using interactive retrieval of social media images | |
CN107291895B (zh) | 一种快速的层次化文档查询方法 | |
Wang et al. | Mining changes of classification by correspondence tracing | |
CN115688760B (zh) | 一种智能化导诊方法、装置、设备及存储介质 | |
CN103778206A (zh) | 一种网络服务资源的提供方法 | |
Yilmaz et al. | Using ontology and sequence information for extracting behavior patterns from web navigation logs | |
JP5014479B2 (ja) | 画像検索装置、画像検索方法及びプログラム | |
CN113762703A (zh) | 确定企业画像的方法和装置、计算设备和存储介质 | |
CN113326432A (zh) | 一种基于决策树与推荐方法的模型优选方法 | |
CN111797267A (zh) | 一种医学图像检索方法及系统、电子设备、存储介质 | |
CN114463072A (zh) | 基于业务需求ai预测的电商服务优化方法及大数据系统 | |
CN114219089A (zh) | 一种新一代信息技术产业知识图谱的构建方法及设备 | |
JP5971722B2 (ja) | ハッシュ関数の変換行列を定める方法、該ハッシュ関数を利用するハッシュ型近似最近傍探索方法、その装置及びそのコンピュータプログラム | |
Schenker et al. | Comparison of distance measures for graph-based clustering of documents |
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: 20927115 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/01/2023) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20927115 Country of ref document: EP Kind code of ref document: A1 |