Method for identifying important assets and key links in network asset map

Abstract

The application discloses a kind of identification method of important asset and key link in network asset diagram, the identification method is first combined with the advantage of random walk betweenness centrality and degree centrality, using random walk betweenness centrality obtains candidate cluster center set, then in candidate cluster center using degree centrality filters small degree node to obtain cluster center set, then the cluster center set is divided into buckets, and subdivides cluster center set, can successfully identify the business influence cluster center in important asset, according to business rule, the bridging link between important asset, that is, key link, is identified on the basis of known cluster center;Second, another important asset-business influence bridge node on key link is obtained by extracting bridging link.The application provides a method for identifying important assets and key links in network asset diagrams, which has the characteristics of good identification effect and high accuracy, and is suitable for popularization and application.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of knowledge graph mining, and particularly relates to a method for identifying important assets and key links in a network asset graph. BACKGROUND

[0002] Graph is a general language for describing and modeling complex systems, and is ubiquitous in the real world. For example, social media such as Facebook and Twitter constitute a social network between humans; protein molecules in the human body constitute a biological network; various mobile terminals constitute a communication network; intelligent hardware constitutes an Internet of Things; highways, railways, and routes between cities constitute a transportation network, etc. With the advent of the big data era, data mining on graphs has gradually become one of the research hotspots, such as recommending friends of interest to users, judging protein structure, predicting traffic flow, and detecting abnormal accounts. These graph data mining algorithms not only improve the visualization of graph structure, but also help us better understand the complex relationships in the network.

[0003] Knowledge graph is a technical method for describing knowledge and modeling the relationship between all things in the world by using a graph model. A knowledge graph is composed of nodes and edges. A node can be an entity, such as a person, a book, etc., or an abstract concept, such as artificial intelligence, knowledge graph, etc. An edge can be an attribute of an entity, such as a name, a book title, or a relationship between entities, such as friends, spouse. The early concept of knowledge graph comes from Semantic Web, which originally aimed to transform the World Wide Web based on text links into a semantic web based on entity links.

[0004] Network asset graph is also a complex knowledge graph and heterogeneous network, which is composed of multiple types of network assets and their associated relationships. Among the important nodes and edges of the network asset graph, the business influence cluster center and the bridge node are called important assets, and the bridge link formed by the important assets and their edges is called key link. These two structures play a very important role in the graph. However, as the size of the network asset graph increases, it is difficult for analysts to directly identify important assets and key links by naked eye. SUMMARY

[0005] In order to solve the defects and deficiencies of the current technology, the purpose of the present application is to provide a method for identifying important assets and key links in a network asset graph. The identification method has high accuracy, good extraction effect, and can improve the visual space.

[0006] The object of the present application is achieved by the following technical solutions:

[0007] A method for identifying important assets and key links in a network asset map, comprising the following steps:

[0008] S01: Obtain a network asset map, calculate the random walk betweenness centrality of each node in the network asset map, and amplify;

[0009] S02: Input the amplified random walk betweenness centrality of each node in step S01 into the first h-index algorithm to obtain T1;

[0010] S03: If the amplified random walk betweenness centrality of a node in step S01 is greater than T1 obtained in step S02, add the node to the candidate cluster center set;

[0011] S04: Calculate the degree centrality of each node in the candidate cluster center set obtained in step S03 and amplify it;

[0012] S05: Input the amplified degree centrality of each node in step S04 into the second h-index algorithm to obtain T2;

[0013] S06: If the amplified degree centrality of a node in step S04 is greater than T2 obtained in step S05, add the node to the cluster center set;

[0014] S07: Perform bucketing on the cluster center set obtained in step S06;

[0015] S08: According to the cluster center set obtained in step S06, identify the business influence cluster center;

[0016] S09: According to the cluster center set obtained in step S06 and the bucketing result of step S07, identify the key link and bridge node;

[0017] S10: According to the set of bridge nodes obtained in step S09, identify the business influence bridge node.

[0018] In one embodiment, in step S02, the specific steps of the first h-index algorithm are as follows:

[0019] S02.1: Sort all nodes according to the amplified random walk betweenness centrality in step S01 from large to small;

[0020] S02.2: Assuming h = i (i = 1, 2, …, n), n is the number of nodes, starting from the node with the largest centrality after sorting, if h * s_1 ≥ the centrality of the hth node, then only h-1 nodes have a centrality greater than (h-1) * s_1, at this time, T_1 = (h-1) * s_1, wherein s_1 is a scaling index predefined by a user.

[0021] In one of the embodiments, in step S05, the specific steps of the second h-index algorithm are as follows:

[0022] S05.1, for all nodes, sort the degree centrality amplified in step S04 from large to small;

[0023] S05.2, assuming h = i (i = 1, 2, …, n), n is the number of nodes, starting from the node with the largest centrality after sorting, if h * s_2 ≥ the centrality of the hth node, then only h-1 nodes have a centrality greater than (h-1) * s_2, at this time, T_2 = (h-1) * s_2, wherein s_2 is a scaling index predefined by a user.

[0024] In one of the embodiments, in step S08, the specific steps of identifying the business influence cluster center are as follows:

[0025] S08.1: Observe the structure and business characteristics of the cluster center on the network asset map;

[0026] S08.2: Divide the cluster center into business influence cluster center, node influence cluster center, and structure influence cluster center;

[0027] S08.3: If a cluster center (node) is a business influence cluster center, the cluster center is considered an important asset.

[0028] In one of the embodiments, in step S09, the specific steps of identifying the key link are as follows:

[0029] S09.1: Traverse all cluster center nodes to find bridge links between different cluster centers, and then output the bridge nodes;

[0030] S09.2: Traverse the bridge link, if both ends of the bridge link are the business influence cluster center, then the bridge link is a key link.

[0031] In one of the embodiments, in step S10, the specific steps of identifying the business influence bridge node are as follows:

[0032] S10.1, observe the structure and business characteristics of the bridge node on the network asset map;

[0033] S10.2, divide the bridge node into a business influence bridge node, a first-level bridge node, a second-level bridge node and a third-level bridge node;

[0034] S10.3, if a bridge node is a business influence bridge node, the bridge node is regarded as an important asset.

[0035] The present application has the following advantages and effects relative to the prior art: the present application provides a method for identifying important assets and key links in a network asset map, which has simple parameter setting and accurate extraction effect, applies the top-down and classified refinement idea to the detection of network assets and asset chains, effectively solves the problems of high dependence on user-defined parameters and low accuracy in existing recognition algorithms, and helps people to deeply mine the special business structure of the network asset map, so that the present application is suitable for popularization and application. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 Fig. 1 is a flowchart of the method for identifying important assets and key links in a network asset map according to an embodiment of the present application;

[0037] Figure 2 Fig. 2 is a network asset map before extraction of case data data7 according to an embodiment of the present application;

[0038] Figure 3 Fig. 3 is a network asset map after extraction of data data7 according to the embodiment of the present application; Figure 2

[0039] Fig. 4 is a network asset map before extraction of case data data14 according to another embodiment of the present application; Figure 4

[0040] Fig. 5 is a network asset map after extraction of data data14 according to the embodiment of the present application. Figure 5 DETAILED DESCRIPTION Figure 4 The present application will be further described in detail below with reference to embodiments, but the embodiments of the present application are not limited thereto.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0042] As used herein, the term "containing" or "including" can be open, semi-closed and closed. In other words, the term also includes "consisting essentially of" or "consisting of".

[0043] As used herein, the term "containing" or "including" can be open, semi-closed and closed. In other words, the term also includes "consisting essentially of" or "consisting of".

[0044] ​The application discloses a method for identifying important assets and key links in a network asset map, also known as an Extract algorithm, which is realized by HubExtract and BridgeExtract algorithms.

[0045] Specifically, as shown in the drawings, Figure 1 The application discloses a method for identifying important assets and key links in a network asset map, also known as an Extract algorithm, which is realized by HubExtract and BridgeExtract algorithms.

[0046] S01: Obtain a network asset map, calculate random walk betweenness centrality of each node in the network asset map, and amplify the random walk betweenness centrality;

[0047] S02: Input the amplified random walk betweenness centrality of each node in step S01 into a first h-index algorithm to obtain T1 (threshold 1);

[0048] S03: If the amplified random walk betweenness centrality of a certain node in step S01 is greater than T1 obtained in step S02, the node is added to a candidate cluster center set;

[0049] S04: Calculate the degree centrality of each node in the candidate cluster center set obtained in step S03 and amplify the degree centrality;

[0050] S05: Input the amplified degree centrality of each node in step S04 into a second h-index algorithm to obtain T2 (threshold 2);

[0051] S06: If the amplified degree centrality of a certain node in step S04 is greater than T2 obtained in step S05, the node is added to a cluster center set;

[0052] S07: Perform bucketing on the cluster center set obtained in step S06;

[0053] S08: identifying the business influential cluster center according to the cluster center set obtained in step S06;

[0054] S09: identifying the key link and bridge node according to the cluster center set obtained in step S06 and the bucketing result in step S07;

[0055] S10: identifying the business influential bridge node according to the set of bridge nodes obtained in step S09.

[0056] In one embodiment, in step S02, the specific steps of the first h-index algorithm are as follows:

[0057] S02.1: sorting all nodes according to the amplified random walk betweenness centrality in step S01 from large to small;

[0058] S02.2: assuming h=i (i=1, 2, …, n), n is the number of nodes, starting from the node with the largest centrality after sorting, if h*s_1≥the centrality of the hth node, then only h-1 nodes have a centrality greater than (h-1)*s_1, at this time, T_1=(h-1)*s_1, wherein s_1 is a user-defined scaling index.

[0059] In one embodiment, in step S05, the specific steps of the second h-index algorithm are as follows:

[0060] S05.1, sorting all nodes according to the amplified degree centrality in step S04 from large to small;

[0061] S05.2, assuming h=i (i=1, 2, …, n), n is the number of nodes, starting from the node with the largest centrality after sorting, if h*s_2≥the centrality of the hth node, then only h-1 nodes have a centrality greater than (h-1)*s_2, at this time, T_2=(h-1)*s_2, wherein s_2 is a user-defined scaling index.

[0062] In one embodiment, in step S08, the specific steps of identifying the business influential cluster center are as follows:

[0063] S08.1: observing the structure and business characteristics of the cluster center on the network asset map;

[0064] S08.2: dividing the cluster center into a business influential cluster center, a node influential cluster center, and a structural influential cluster center;

[0065] S08.3: if a cluster center (node) is a business influential cluster center, the cluster center is considered an important asset.

[0066] In one embodiment, in step S09, the specific steps of identifying the key link are as follows:

[0067] S09.1: traverse all cluster center nodes, find the bridging link between different cluster centers, and then output the bridge node;

[0068] S09.2: traverse the bridging link, if both ends of the bridging link are the service influence cluster center, the bridging link is a key link.

[0069] In one embodiment, in step S10, the specific steps of identifying the service influence bridge node are as follows:

[0070] S10.1, observe the structure and service characteristics of the bridge node on the network asset map;

[0071] S10.2, divide the bridge node into a service influence bridge node, a first-level bridge node, a second-level bridge node, and a third-level bridge node;

[0072] S10.3, if a bridge node is a service influence bridge node, the bridge node is considered as an important asset.

[0073] In one embodiment, the identification method of the present application is used to identify the case data data7, and the network asset map before extraction of the case data data7 is as shown in Figure 2 , and the network asset map after extraction of the data data7 is as shown in Figure 3 . As can be seen from Figure 2 and Figure 3 , the important assets and key links in the network asset map can be effectively identified by the identification method of the present application.

[0074] In another embodiment, the identification method of the present application is used to identify the case data data14, and the network asset map before extraction of the case data data14 is as shown in Figure 4 , and the network asset map after extraction of the data data14 is as shown in Figure 5 . As can be seen from Figure 4 and Figure 5 , the important assets and key links in the network asset map can be effectively identified by the identification method of the present application.

[0075] The present application provides a method for identifying important assets and key links in a network asset map, which has simple parameter setting and accurate extraction effect. The top-down and classification refinement idea is applied to the detection of network assets and asset chains, effectively solving the problem of high dependence on user-defined parameters and low accuracy in existing identification algorithms, helping people to deeply mine the special business structure of the network asset map, and being suitable for popularization and application.

[0076] The above embodiments are the preferred embodiments of the present application, but the embodiments of the present application are not limited to the above embodiments, and any changes, modifications, substitutions, combinations, simplifications, etc. made without departing from the spirit and principles of the present application should be equivalent replacement manners and should be included in the protection scope of the present application.

Claims

1. A method for identifying important assets and critical links in a network asset graph, characterized in that, Includes the following steps: S01: Obtain the network asset graph, calculate the random walk betweenness centrality of each node in the network asset graph, and zoom in; S02: Input the magnified random walk betweenness centrality of each node in step S01 into the first h-index algorithm to obtain T1; S03: If the random walk betweenness centrality of a node after being amplified in step S01 is greater than T1 obtained in step S02, then add the node to the candidate cluster center set. S04: Calculate the degree centrality of each node in the candidate cluster center set obtained in step S03, and amplify it; S05: Input the amplified degree centrality of each node in step S04 into the second h-index algorithm to obtain T2; S06: If the degree centrality of a node after being amplified in step S04 is greater than T2 obtained in step S05, then add the node to the cluster center set. S07: Bucket the cluster center set obtained in step S06; S08: Identify the business influence cluster centers based on the cluster center set obtained in step S06; S09: Based on the cluster center set obtained in step S06 and the bucketing results in step S07, identify the critical links and bridge nodes; S10: Based on the set of bridge nodes obtained in step S09, identify the bridge nodes with business influence; In step S09, the specific steps for identifying the critical link are as follows: S09.1: Traverse all cluster center nodes, find the bridging links between different cluster centers, and then output the bridge nodes; S09.2: Traverse the bridging links. If both ends of a bridging link are the business influence cluster center, then the bridging link is a critical link. In step S08, the specific steps for identifying the center of the business influence cluster are as follows: S08.1: Observe the structure and business characteristics of the cluster center on the network asset graph; S08.2: Divide the cluster centers into business influence cluster centers, node influence cluster centers, and structural influence cluster centers; S08.3: If a cluster center is a business-influence cluster center, then the cluster center is considered an important asset; In step S10, the specific steps for identifying the business influence bridge node are as follows: S10.1 Observe the structure and business characteristics of the bridge nodes on the network asset graph; S10.2 Divide the bridge nodes into business influence bridge nodes, first-level bridge nodes, second-level bridge nodes, and third-level bridge nodes; S10.3 If a bridge node is a bridge node with business influence, then the bridge node is considered an important asset.

2. The method for identifying important assets and critical links in a network asset graph according to claim 1, characterized in that, In step S02, the specific steps of the first h-index algorithm are as follows: S02.1: Sort all nodes in descending order according to the random walk betweenness centrality amplified in step S01; S02.2: Assume h = i (i = 1, 2, ..., n), where n is the number of nodes. Starting from the node with the largest centrality after sorting, traverse downwards. If h * s_1 ≥ the centrality of the h-th node, then only the centrality of h-1 nodes is greater than (h-1) * s_1. In this case, T1 = (h-1) * s_1, where s_1 is a user-defined scaling factor.

3. The method for identifying important assets and critical links in a network asset graph according to claim 1, characterized in that, In step S05, the specific steps of the second h-index algorithm are as follows: S05.

1. Sort all nodes in descending order according to the degree centrality amplified in step S04. S05.2 Assume h = i (i = 1, 2, ..., n), where n is the number of nodes. Starting from the node with the largest centrality after sorting, traverse downwards. If h * s_2 ≥ the centrality of the h-th node, then only the centrality of h-1 nodes is greater than (h-1) * s_2. In this case, T2 = (h-1) * s_2, where s_2 is a user-defined scaling factor.

Images