Marine transportation network security evaluation method and model based on hybrid weight and inverse distance center aggregation

Abstract

The invention discloses a sea transportation network security evaluation method and model based on hybrid weight and inverse distance center aggregation, and the method comprises the steps: obtaining network data and security impact factors of a sea transportation network, calculating the importance of each node in the sea transportation network, and calculating each network evaluation index in a multi-level evaluation system; constructing a security level division reference of the network, and dividing the network into a high security level, a middle security level and a low security level; screening out independent evaluation indexes based on the variance expansion factor, carrying out standardization processing, and constructing a standardized data set; respectively processing the standardized data set by using variance analysis and a random forest model based on the divided security levels, and carrying out weighted fusion on processing results to obtain comprehensive weights of the independent evaluation indexes; calculating a dynamic safety score of each independent evaluation index by adopting an inverse distance dynamic mapping mechanism based on the divided safety level and the standardized data set; and constructing a weighted aggregation model to summarize the dynamic security score and the mixed weight of each independent index, and outputting the overall security score of the marine transportation network.

Description

Technical Field

[0001] The present invention relates to the technical field of geospatial information analysis and multi-source data fusion, and particularly relates to a method and model for evaluating the security of a maritime transportation network based on hybrid weights and inverse-distance center aggregation. Background Art

[0002] At present, research on maritime transportation safety mainly focuses on port safety and sea route safety. Existing achievements are mostly limited to the isolated analysis of local nodes or single channels. Moreover, scholars' analysis of the overall security of the maritime transportation network mainly focuses on certain single reliability indicators, lacking a systematic evaluation of the security of the maritime transportation network and its multi-dimensional security coupling mechanism from an overall perspective, making it difficult to support the overall security optimization requirements of the ocean transportation network in a complex environment. Currently, when constructing an evaluation model for the overall security of the maritime transportation network, methods such as the analytic hierarchy process and Bayesian networks are mostly used. The construction of the model requires a large amount of prior knowledge support and relies on expert experience, with strong subjectivity; or through objective weighting methods such as the entropy weight method and coefficient of variation method, the construction of the model overly relies on data distribution and characteristics, which easily leads to weight deviation and poor interpretability.

[0003] In summary, the existing maritime transportation network evaluation models have not solved the problem of global security evaluation with multi-dimensional index fusion, and it is difficult to meet the dynamic risk assessment requirements in a complex environment. There is an urgent need to develop a new evaluation framework that integrates the advantages of subjective and objective weighting, dynamic weight adjustment, and multi-source data collaboration to achieve precise, dynamic, and systematic quantitative analysis of the security of the maritime transportation network. Summary of the Invention

[0004] To overcome the deficiencies that the existing maritime transportation network evaluation models in the prior art are difficult to perform global security evaluation with multi-dimensional index fusion and are difficult to adapt to dynamic risk assessment in a complex environment, the present invention provides a method and model for evaluating the security of a maritime transportation network based on hybrid weights and inverse-distance center aggregation. Through a multi-level evaluation system and the division of network security levels, key indicators are screened and weighted by comprehensively verifying the significance of variance analysis and analyzing with a random forest model. Combining the inverse-distance dynamic mapping mechanism between the indicator values and the level center, systematic quantitative evaluation of the security of the maritime transportation network is achieved through weighted aggregation.

[0005] According to one aspect of the specification of the present invention, there is provided a method for evaluating the security of a maritime transportation network based on hybrid weights and inverse-distance center aggregation, including:

[0006] Obtaining the network data and security impact factors of the maritime transportation network, calculating the importance of each node in the maritime transportation network, and then calculating each network evaluation index in the multi-level evaluation system;

[0007] Construct a multi-level evaluation system for the maritime transportation network, obtain the network data and security impact factors of the maritime transportation network, calculate the importance of each node in the maritime transportation network, and then calculate each network evaluation index in the multi-level evaluation system;

[0008] Construct a benchmark for classifying the security level of the network and classify the security level of the network;

[0009] Screen out independent evaluation indicators based on the variance inflation factor and perform standardization processing, group them based on the security level of the network, and then construct a standardized data set;

[0010] Process the standardized data set using analysis of variance and the random forest model respectively, and fuse the processing results weighted to obtain the comprehensive weights of each independent evaluation indicator;

[0011] Calculate the dynamic security scores of each independent evaluation indicator based on the standardized data set using the inverse distance dynamic mapping mechanism;

[0012] Construct a weighted aggregation model to summarize the dynamic security scores and mixed weights of each independent indicator, and output the overall security score of the maritime transportation network.

[0013] As a further implementation plan, the calculation steps of the node importance are as follows:

[0014] Calculate the normalized degree centrality, normalized closeness centrality, and normalized betweenness centrality evaluation indicators of each node based on the topological structure of the network and the historical monthly OD connection quantity of the nodes. Sort all the nodes of the maritime transportation network in different central point dimensions according to the calculation results, and calculate the scores of each node in different central point dimensions according to the sorting situation;

[0015] Fuse the scores of each node in different central point dimensions weighted to calculate the comprehensive centrality score of the node, and normalize the sum of the comprehensive scores of all nodes in the network as the node importance.

[0016] As a further implementation plan, the multi-level evaluation system includes four first-level security core evaluation dimensions, including material security, topological structure security, dynamic evolution security, and geopolitical environment security;

[0017] Material security includes four second-level evaluation indicators, namely import stability, export stability, trade dependence, and material concentration;

[0018] Topological structure security includes five second-level evaluation indicators, namely modularity, ratio of the largest connected subgraph, network clustering coefficient, network average degree, and network density;

[0019] Dynamic evolution security includes seven secondary evaluation indicators, namely node dynamic evolution stability, edge dynamic evolution stability, modularity change ratio, maximum connected subgraph ratio change ratio, network clustering coefficient change ratio, network average degree change ratio, and network density change ratio;

[0020] Geographical environment security includes six secondary evaluation indicators, namely network trade balance value, earthquake disturbance, sea ice disturbance, sea surface height, geopolitical risk, and consumer price.

[0021] As a further implementation plan, network data includes the monthly OD connection quantity and monthly OD trade quantity of shipping history;

[0022] Security impact factors include, but are not limited to, earthquake disturbance, sea ice disturbance, geopolitical risk, consumer price, and network trade balance value of nodes in the network.

[0023] As a further implementation plan, the benchmark for classifying the security level of the shipping network includes two aspects: functional security and structural security. Specifically:

[0024] Establish a network security level classification benchmark corresponding to functional security and structural security respectively, and divide the shipping network into three security levels: 0, 1, and 2;

[0025] Based on the structural security level and functional security level of the integrated shipping network, classify the overall security level of the shipping network. The classification rules are as follows:

[0026]

[0027] Among them, 、 represent the structural security level and functional security level of the shipping network respectively.

[0028] As a further implementation plan, the specific process of screening based on the variance inflation factor is as follows:

[0029] Construct a linear regression model corresponding to each evaluation indicator of the network, calculate the variance inflation factor, set the maximum threshold of the variance inflation factor for data cleaning, and retain the evaluation indicators as independent evaluation indicators.

[0030] As a further implementation plan, the process of obtaining the comprehensive weight of the evaluation indicators of independent evaluation indicators is specifically as follows:

[0031] For the standardized data set, calculate the within-group variance and between-group variance of different network security level groups for each independent evaluation indicator respectively, and then use the within-group variance and between-group variance to calculate the F statistic and p value, and use multiple test correction of the p value to avoid false positives. Finally, select the independent evaluation indicators with statistical significance and output the corrected p value of the indicators;

[0032] Input the standardized data set to train the random forest model. The trained model outputs the feature importance scores of independent evaluation indicators based on the Gini importance of features for model credibility verification, and at the same time outputs the permutation importance scores of independent evaluation indicators.

[0033] Weightedly fuse the corrected p-values of independent evaluation indicators with statistical significance and the permutation importance scores of independent evaluation indicators to obtain the comprehensive weights of independent evaluation indicators.

[0034] As a further implementation, the steps of calculating the dynamic security scores of each independent evaluation indicator using the inverse distance dynamic mapping mechanism are specifically as follows:

[0035] For the standardized data set, for different network security level groups, calculate the standardized mean center of each evaluation indicator within the group respectively.

[0036] Calculate the Euclidean distance from the evaluation indicators within the group to the standardized mean center according to the standardized mean center, and combine the smoothing coefficient to calculate the normalized weight of the evaluation indicators for the network security level.

[0037] Construct an inverse distance weight dynamic scoring model, and combine the normalized weights of the evaluation indicators at different security levels to output the dynamic security scores of the evaluation indicators. As a further implementation, the mathematical representation of the weighted aggregation model is as follows:

[0038]

[0039] Among them, is the dynamic security score of the independent evaluation indicator of the network ; is the comprehensive weight of the independent evaluation indicator ; is the comprehensive dynamic security score of the network ; is the total number of independent evaluation indicators.

[0040] According to one aspect of the specification of the present invention, there is provided a maritime network security evaluation model based on hybrid weight and inverse distance center aggregation, including:

[0041] A data collection and processing module, responsible for collecting the network data of the maritime network and the security impact factors, and calculating the importance of each node and various network evaluation indicators.

[0042] A security level division module, used to divide the security level of the network.

[0043] A standardized data set construction module, which screens independent evaluation indicators and constructs a standard data set.

[0044] A comprehensive weight calculation module that obtains the comprehensive weight of independent evaluation indicators;

[0045] A dynamic scoring module that obtains the dynamic security scores of independent evaluation indicators;

[0046] A security scoring output module that outputs the overall security score of the maritime transportation network.

[0047] Compared with the prior art, the beneficial effects of the present invention are as follows: Through a multi-level evaluation system and the division of network security levels, the present invention comprehensively uses variance analysis significance verification and random forest model analysis to screen key indicators and assign weights, combines the inverse distance dynamic mapping mechanism between the indicator values and the level centers, and realizes the systematic quantitative evaluation of maritime transportation network security through weighted aggregation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings used in the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0049] Figure 1 It is a schematic flow chart of a method for evaluating the security of a strategic material maritime transportation network based on a hybrid weight and inverse distance center aggregation in an embodiment of the present invention;

[0050] Figure 2 It is a schematic diagram of a multi-level evaluation system in an embodiment of the present invention;

[0051] Figure 3 It is a schematic diagram of the division of maritime transportation network security levels in an embodiment of the present invention;

[0052] Figure 4 It is a schematic diagram of the calculation results of network security levels in an embodiment of the present invention;

[0053] Figure 5 It is a schematic diagram of the results of network dynamic security scores in an embodiment of the present invention;

[0054] Figure 6 It is a schematic diagram of a model for evaluating the security of a strategic material maritime transportation network based on a hybrid weight and inverse distance center aggregation in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] It should be noted that:

[0056] In the description, claims and the above-mentioned drawings of the present invention, the terms "comprising", "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0057] The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices. The flowcharts shown in the drawings are only illustrative and do not have to include all the content and operations / steps, nor do they have to be executed in the described order. For example, some operations / steps can be decomposed, while some operations / steps can be combined or partially combined, so the actual execution order may change according to the actual situation.

[0058] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. In addition, the technical features in each embodiment or a single embodiment provided by the present invention can be combined with each other arbitrarily to form a new technical solution. This combination is not restricted by the order of steps and / or the structural composition mode, but must be based on the fact that those of ordinary skill in the art can implement it. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0059] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a strategic material sea transportation network security evaluation method based on hybrid weight and inverse distance center aggregation in an embodiment of the present invention, and includes:

[0060] Obtain the network data and security impact factors of the sea transportation network, calculate the importance of each node in the sea transportation network, and then calculate each network evaluation index in the multi-level evaluation system;

[0061] Construct a multi-level evaluation system for the sea transportation network and obtain the network data and security impact factors of the sea transportation network, calculate the importance of each node in the sea transportation network, and then calculate each network evaluation index in the multi-level evaluation system;

[0062] Construct a security level division criterion for the network and divide the security levels of the network;

[0063] Screen out independent evaluation indicators based on the variance inflation factor and perform standardization processing. Group them based on the security levels of the network, and then construct a standardized data set;

[0064] Process the standardized data set using analysis of variance and the random forest model respectively, and the processing results are weighted and fused to obtain the comprehensive weights of each independent evaluation indicator;

[0065] Based on the standardized data set, calculate the dynamic security scores of each independent evaluation indicator using the inverse distance dynamic mapping mechanism;

[0066] Construct a weighted aggregation model to summarize the dynamic security scores and hybrid weights of each independent indicator, and output the overall security score of the maritime network.

[0067] Furthermore, please refer to Figure 2 , Figure 2 which is a schematic diagram of the multi-level evaluation system in the embodiment of the present invention. The multi-level evaluation system includes four first-level security core evaluation dimensions, including material security, topological structure security, dynamic evolution security, and geopolitical environment security;

[0068] Material security includes four secondary evaluation indicators, namely import stability, export stability, trade dependence, and material concentration;

[0069] Topological structure security includes five secondary evaluation indicators, namely modularity, ratio of the largest connected subgraph, network clustering coefficient, network average degree, and network density;

[0070] Dynamic evolution security includes seven secondary evaluation indicators, namely node dynamic evolution stability, edge dynamic evolution stability, modularity change ratio, ratio of the largest connected subgraph change ratio, network clustering coefficient change ratio, network average degree change ratio, and network density change ratio;

[0071] Geopolitical environment security includes six secondary evaluation indicators, namely network trade balance value, earthquake disturbance, sea ice disturbance, sea surface height, geopolitical risk, and consumer price.

[0072] Furthermore, the network data includes the historical monthly OD connection quantity and the historical monthly OD trade quantity of maritime transportation;

[0073] The security impact factors include, but are not limited to, earthquake disturbance, sea ice disturbance, geopolitical risk, consumer price, and network trade balance value of the nodes in the network.

[0074] Furthermore, the calculation steps of the node importance are as follows:

[0075] Calculate the normalized degree centrality, normalized closeness centrality, and normalized betweenness centrality evaluation indicators for each node based on the network topology and the historical monthly OD connection volume of the nodes. Sort all the nodes of the maritime network according to the calculation results in different central point dimensions, and calculate the scores of each node in different central point dimensions according to the sorting situation.

[0076] Weightedly fuse the scores of each node in different central point dimensions to calculate the comprehensive centrality score of the node, and normalize it to obtain the sum of the comprehensive scores of all nodes in the network as the node importance.

[0077] Specifically, input the historical monthly OD connection volume and trade volume data, calculate the node sorting situation obtained according to different centrality evaluation indicators for the importance of each node at each time, and calculate the score of each node:

[0078]

[0079]

[0080]

[0081] In the formula, represents the total number of nodes in the network; , and respectively represent the rankings of node in the sample network based on the normalized degree centrality, normalized closeness centrality, and normalized betweenness centrality; , and respectively represent the scores of this node based on the ranking of this centrality.

[0082] Based on the above scores, calculate the comprehensive centrality score of this node:

[0083]

[0084] Among them, represents the comprehensive score of node ; , and respectively represent the weights of each score in the comprehensive score of the node. This weight is determined by the correlation between the normalized degree centrality, closeness centrality, and betweenness centrality of the node and the throughput of each strategic material of this node :

[0085]

[0086]

[0087] , and are the correlation coefficients between each centrality and the node material throughput respectively.

[0088] Normalize the comprehensive score of node :

[0089]

[0090] is the comprehensive importance score of the normalized node ; is the sum of the comprehensive scores of all nodes in the network; is the total number of network nodes.

[0091] Specifically, how to calculate the calculation formulas and various indicators of each network evaluation index of the network by combining the historical monthly OD connection quantity and trade volume data and safety impact factor data through the node importance is explained as follows:

[0092] The import stability CV is the ability of the network material import node to quickly obtain new supplies, and the calculation formula is as follows:

[0093]

[0094] In the formula, represents the stability of a certain material importing country ; represents the number of all the material exporting countries in the network; represents all the voyages between the export port i and the import port j; represents the total export volume of the material exporting country ; represents the total world export volume; represents the number of all the material importing countries; represents the shortest path length between the material exporting country and the material importing country . When there is no shortest path length between and , .

[0095] The export stability ES is the ability of the network material export node to maintain export stability, and the calculation formula is as follows:

[0096]

[0097] In the formula, represents the stability of a certain material exporting country ; Indicates the number of all importing countries of this material in the network; Represents all voyages between import port l and export port k; Indicates the importing country of this material Total import amount; Indicates the total world import amount; Represents the number of all exporting countries of this material; Indicates the importing country of this material And the exporting country of this material The shortest path length between them. When And There is no shortest path length between them, .

[0098]

[0099] In the formula, ES represents the overall export stability of the seaborne network of this material; Is the comprehensive importance score of node i after normalization.

[0100] The trade dependence degree TDA reflects the dependence degree of network trade on a single trade relationship. The calculation formula is:

[0101]

[0102] In the formula, TDA represents the trade dependence degree of the seaborne network of this material; Is the trade volume between the largest nodes in the network trade volume; Is the total network trade volume.

[0103] The material concentration degree CCC reflects the dependence degree of network trade on a single trade relationship. The calculation formula is:

[0104]

[0105] In the formula, CCC represents the material concentration degree of the seaborne network of this material; Is the trade volume processed by the largest node in the network; Is the total network trade volume.

[0106] The modularity Q is used to measure the quality of network community division. The calculation formula is:

[0107]

[0108] In the formula, Q is the modularity of the seaborne network of the material; Is the edge weight between node i and node j; And Are the degrees of node i and node j respectively; E is the total number of edges in the network; It is an indicator function that takes the value of 1 when nodes i and j belong to the same community and 0 otherwise.

[0109] The ratio S of the largest connected subgraph is the ratio of the number of nodes in the largest connected subgraph to the total number of nodes in the network, and the calculation formula is:

[0110]

[0111] In the formula, S is the ratio of the largest connected subgraph of the material sea transportation network; represents the number of nodes in the largest connected subgraph; n represents the total number of nodes in the network.

[0112] The network clustering coefficient C is the ratio of the number of connections between all adjacent nodes to the possible maximum number of connections, and the calculation formula is:

[0113]

[0114] In the formula, C is the network clustering coefficient; is the degree of node i; is the actual number of connection edges between node i and its adjacent nodes; n represents the total number of nodes in the network.

[0115] The network average degree K is the average value of the number of connection edges between all the same nodes in the network, and the calculation formula is:

[0116]

[0117] In the formula, K is the average degree of the sea transportation network; n is the number of nodes in the network; is the degree of node i.

[0118] The network density P is the ratio of the number of actually existing edges in the network to the maximum number of edges that could theoretically exist, and the calculation formula is:

[0119]

[0120] In the formula, P is the network density; E is the number of actually existing edges; n is the number of actually existing nodes.

[0121] Node dynamic evolution stability It is used to measure the change in the number of countries between two networks from month t to month t + 1, and the calculation formula is:

[0122]

[0123] In the formula, is the node dynamic evolution stability; is the set of countries that joined the international trade of this material in year t; is the number of common countries in year t and year t + 1; is the number of all countries in year t and year t + 1.

[0124] Dynamic Evolutionary Stability of Edges It is used to measure the change in the trade relationship between two networks from month t to month t + 1, and the calculation formula is:

[0125]

[0126] In the formula, is the dynamic evolutionary stability of the edge; is the set of international trade relationships of a certain material in month t; is the number of common trade relationships in month t and month t + 1; is the number of all trade relationships in month t and month t + 1.

[0127] Ratio of Modularity Change It represents the ratio of the change in modularity between two networks from month t to month t + 1, and the calculation formula is:

[0128]

[0129] In the formula, is the ratio of modularity change; Q and are the modularities in month t and month t + 1 respectively.

[0130] Ratio of Change in the Largest Connected Subgraph Ratio It represents the ratio of the change in the largest connected subgraph ratio between two networks from month t to month t + 1, and the calculation formula is:

[0131]

[0132] In the formula, is the ratio of change in the largest connected subgraph ratio; S and are the largest connected subgraph ratios in month t and month t + 1 respectively.

[0133] Ratio of Change in Network Clustering Coefficient , which represents the ratio of the change in the clustering coefficient between two networks from month t to month t + 1, and the calculation formula is:

[0134]

[0135] In the formula, is the ratio of change in the network clustering coefficient; C and are the clustering coefficients in month t and month t + 1 respectively.

[0136] Ratio of Change in Network Average Degree It represents the ratio of the change in the network average degree between two networks from month t to month t + 1, and the calculation formula is:

[0137]

[0138] wherein, is the change ratio of the network average degree; K and are the network average degrees in month t and month t + 1 respectively.

[0139] Change ratio of network density , which represents the ratio of the change in network density between two networks from month t to month t + 1, and the calculation formula is:

[0140]

[0141] wherein, is the change ratio of network density; P and are the network densities in month t and month t + 1 respectively.

[0142] The network trade balance value TBG. Trade balance means that the total value of a country's foreign trade imports and exports basically tends to be balanced in a specific year. The trade balance value refers to the size of the trade volume when trade balance is achieved, and the calculation formula is:

[0143]

[0144] wherein, TBG represents the comprehensive trade balance value of the material network; is the trade balance value of node i, and the data is from the International Monetary Fund; n represents the number of all nodes in the network; is the comprehensive importance score of node i after normalization.

[0145] The earthquake disturbance EQ reflects the overall risk level faced by the network under earthquake disasters, and the calculation formula is:

[0146]

[0147] wherein, EQ represents the comprehensive earthquake disturbance of the material network; is the earthquake disturbance of node i, and the data is from the United States Geological Survey; n represents the number of all nodes in the network; is the node after normalization 's comprehensive importance score.

[0148] The sea ice disturbance SI reflects the overall risk level faced by the network under the change of sea ice environment, and the calculation formula is:

[0149]

[0150] wherein, SI represents the comprehensive sea ice disturbance of the material network; SI is the sea ice perturbation of node i, and the data is from the EUMETSAT OSISAF sea ice concentration dataset and the NOAA / NCDC L4 GHRSST DailyOI SST product; n represents the total number of nodes in the network; is the comprehensive importance score of node i after normalization.

[0151] The sea surface height SH reflects the systematic risk of the network caused by sea level fluctuations, and the calculation formula is:

[0152]

[0153] In the formula, SH represents the comprehensive sea surface height of the material network; SH is the sea surface height of node i, and the data is from the HYCOM Consortium; n represents the total number of nodes in the network; is the comprehensive importance score of node i after normalization.

[0154] The geopolitical risk GP reflects the political tension and uncertainty of the network, and the calculation formula is:

[0155]

[0156] In the formula, GP represents the comprehensive geopolitical risk of the material network; GP is the geopolitical risk of node i, and the data is from the statistical website; n represents the total number of nodes in the network; is the comprehensive importance score of node i after normalization.

[0157] The consumer price CP reflects the systematic risk of the network caused by consumer price fluctuations, and the calculation formula is:

[0158]

[0159] In the formula, CP represents the comprehensive consumer price of the material network; CP is the consumer price of node i, and the data is from the International Monetary Fund; n represents the total number of nodes in the network; is the comprehensive importance score of node i after normalization.

[0160] Furthermore, the safety level division criterion of the maritime transportation network includes two aspects: functional safety and structural safety. Specifically:

[0161] Establish the network security level division criterion for functional safety and structural safety respectively, and divide the maritime transportation network into three safety levels: 0, 1, and 2;

[0162] As Figure 3 shown, Figure 3It is a schematic diagram for the classification of the maritime network security level in the embodiments of the present invention. By integrating the structural security level and the functional security level of the maritime network, the overall security level of the maritime network is classified, and the classification rules are as follows:

[0163]

[0164] Among them, 、 represent the structural security level and the functional security level of the maritime network respectively.

[0165] Specifically, structural security refers to the ability of the network system to ensure normal operation when damaged, which is measured by network resilience, that is, the product term of natural connectivity and network efficiency:

[0166]

[0167] is the network resilience; is the total number of nodes in the network; is the eigenvalue spectrum of the adjacency matrix of the th eigenvalue; is the shortest path length between node and . If the nodes are not connected, , and at this time .

[0168] Based on the mean value and the standard deviation of the data, the classification of the structural security level is carried out:

[0169]

[0170] Functional security refers to the characteristic that the maritime network can still maintain its core operation ability when facing complex environments and disturbances. The total monthly network throughput is used as the core quantitative evaluation index for functional security:

[0171]

[0172] is the total throughput; is the cargo volume of the th node; is the time window, that is, one month.

[0173] Based on the mean value and the standard deviation of the data, the classification of the functional security level is carried out:

[0174]

[0175] Furthermore, the specific process of screening based on the variance inflation factor is as follows:

[0176] For each evaluation index of the network, construct a linear regression model, calculate the variance inflation factor, set the maximum threshold of the variance inflation factor for data cleaning, and the retained evaluation indexes are used as independent evaluation indexes.

[0177] Specifically, detect the multicollinearity between evaluation indexes through the variance inflation factor (VIF), and exclude redundant evaluation indexes with high correlation. For each evaluation index , construct a linear regression model, with as the dependent variable and other evaluation indexes as independent variables, and calculate its variance inflation factor:

[0178]

[0179] is the coefficient of determination of the regression model, reflecting the explanatory degree of other evaluation indexes to . If an evaluation index , it indicates that it is highly collinear with other evaluation indexes and needs to be excluded to ensure data independence.

[0180] Output the set of evaluation indexes after cleaning, and perform Z-score standardization on the retained evaluation indexes to eliminate the influence of dimension differences on subsequent analysis. The standardization formula is:

[0181]

[0182] is the value of the evaluation index of the sample network in the original data ; is the mean value of the evaluation index ; is the standard deviation of the evaluation index ; is the standardized value of the evaluation index of the sample network .

[0183] Generate a standardized data set , which is used for subsequent model training and weight calculation.

[0184] Furthermore, the process of obtaining the comprehensive weight of the evaluation indexes of the independent evaluation indexes is specifically as follows:

[0185] For the standardized dataset, calculate the within-group variance and between-group variance of different network security level groups for each independent evaluation metric. Then, use the within-group variance and between-group variance to calculate the F statistic and p-value, and use multiple testing correction for the p-value to avoid false positives. Finally, select the independent evaluation metrics with statistical significance and output the corrected p-values of the metrics.

[0186] Specifically, analysis of variance (ANOVA) is a statistical method used to test whether there are significant differences in the means among three or more independent groups. Based on the data after removing collinearity, analyze the differences in each evaluation metric under different security levels (low, medium, high) through ANOVA, and select the evaluation metrics that are statistically significant for security level classification.

[0187] For each evaluation metric Perform the following operations:

[0188] Divide the sample networks into three groups according to the security level (low, medium, high): low security ( ), medium security ( ), high security ( ); the number of sample networks in each group is , the total number of sample networks ; the evaluation metric value of the rd sample network in the th group is , and the corresponding evaluation metric value of the th group is .

[0189] Calculate the between-group and within-group variances:

[0190] The total mean represents the average of the evaluation metric values of all sample networks, and the formula is:

[0191]

[0192] represents the total mean; represents the total number of sample networks; represents the th sample network in the th group; is the number of sample networks in the th group.

[0193] The between-group variance (Sum of Squares Between groups, SSB) measures the dispersion of the means among different security level groups, and the formula is:

[0194]

[0195]

[0196] is the mean of the th group; represents the evaluation index value of the th sample network in the th group; is the number of sample networks in the th group;

[0197] The within-group variance (Sum of Squares Within groups, SSW) measures the dispersion of the sample network values within the same security level group. The formula is:

[0198]

[0199] is the within-group sum of squares; represents the evaluation index value of the th sample network in the th group; is the number of sample networks in the th group; is the mean of the

[0200] Calculate the statistic and the value. The statistic is used to test whether the between-group differences are significant. The formula is:

[0201]

[0202] where is the number of groups (here ); represents the total number of sample networks; is the between-group sum of squares; is the within-group sum of squares; The value is calculated according to the distribution table or statistical software, representing the probability that the null hypothesis (the means of all groups are equal) holds. If

[0203] , reject the null hypothesis and conclude that there are significant differences in this evaluation index among different security levels.

[0204]

[0205] The value after Bonferroni correction ; is the total number of evaluation indicators; if , it is considered that the evaluation indicator is significant.

[0206] The standardized dataset is input to train a random forest model. The trained model outputs the feature importance scores of independent evaluation indicators based on the Gini importance of features for model credibility verification, and at the same time outputs the permutation importance scores of independent evaluation indicators.

[0207] Specifically, the standardized dataset is used to quantify the contribution of each indicator to the safety level classification through the Gini importance (GiniImportance) and permutation importance (Permutation Importance) of the random forest model, and to screen key safety evaluation indicators.

[0208] Let the training set , where: is the feature vector of the th sample, corresponding to the set of all independent evaluation indicators of the i-th sample network, , corresponding to -dimensional feature vector, and each dimension represents an independent evaluation indicator; , corresponding to the safety level label, low safety , medium safety , high safety ;

[0209] The random forest consists of decision trees, and the generation of each tree satisfies: bootstrap sampling, sampling samples with replacement from ; dimension-aware feature selection, when splitting nodes, selecting candidate features from the dimension to which it belongs (that is, taking the ceiling of the square root of the total number of features ).

[0210] Calculate the Gini impurity. For the sample set in node , the impurity of the safety level distribution is defined as:

[0211]

[0212] where is the total number of samples in node ; is the number of samples of the th class of safety level in node .

[0213] Calculate the feature importance, the importance score of the feature is calculated as: Calculate as:

[0214]

[0215] where the split gain:

[0216]

[0217] where, is the total number of decision trees in the random forest; is the set of nodes used for splitting in the th tree; , are the number of samples in the left and right nodes after splitting; is the number of samples in the training set.

[0218] Calculate the permutation importance. By randomly shuffling the values of the evaluation metrics and observing the decrease in the model's accuracy, the importance of the evaluation metric is reflected.

[0219] Calculate the accuracy of the original model on the unperturbed data:

[0220]

[0221] is the accuracy of the original model; is the number of samples correctly predicted by the model; is the total number of samples.

[0222] Randomly rearrange the values of the feature to generate the th perturbed dataset :

[0223]

[0224] is the feature vector of the th sample, and its feature value is randomly permuted while other feature values remain unchanged; corresponds to the safety level label.

[0225] The accuracy on the perturbed dataset is calculated as:

[0226]

[0227] is the trained random forest classifier; It is an indicator function, where 1 indicates a correct prediction and 0 otherwise; Corresponding to the security level label; It is the accuracy rate after perturbation; It is the total number of samples.

[0228] Calculate the permutation importance. For a feature The permutation importance is:

[0229]

[0230] It is the accuracy rate of the original model; The accuracy rate after perturbation; For a feature The permutation importance; It is the number of permutation repetitions.

[0231] Weightedly fuse the corrected p-values of statistically significant independent evaluation indicators and the permutation importance scores of independent evaluation indicators to obtain the comprehensive weight of evaluation indicators for independent evaluation indicators.

[0232] Specifically, for an indicator The comprehensive weight Is calculated as:

[0233]

[0234] Is the value after Bonferroni correction; It is the mean of permutation importance; It is the total number of indicators.

[0235] Furthermore, the steps to calculate the dynamic security scores of each independent evaluation indicator using the inverse distance dynamic mapping mechanism are specifically as follows:

[0236] For the standardized dataset, for different network security level groups, calculate the standardized mean center of each evaluation indicator within the group respectively;

[0237] Calculate the Euclidean distance from the evaluation indicator within the group to the standardized mean center according to the standardized mean center, and combine with the smoothing coefficient to calculate the normalized weight of the evaluation indicator for the network security level;

[0238] Construct an inverse distance weight dynamic scoring model, and combine the normalized weights of the evaluation indicator in the node groups of different security levels to output the dynamic security score of the evaluation indicator.

[0239] Specifically, for each security level , calculate the standardized mean center of each evaluation indicator:

[0240]

[0241] is the sample index value after standardization; belongs to the grade number of samples; is the safety level index under standardized mean center.

[0242] The Euclidean distance calculation formula is as follows:

[0243]

[0244] represents the sample index to the grade Euclidean distance to the center; is the sample index value after standardization; is the safety level index under standardized mean center.

[0245] The weight calculation formula is as follows:

[0246]

[0247] , is the smoothing coefficient to prevent the denominator from being 0 due to zero distance; represents the sample index to the grade Euclidean distance to the center; is the index for the grade normalized weight.

[0248] For the sample index , its dynamic safety score is calculated as:

[0249]

[0250] is the grade benchmark score ( ); is the index for the grade normalized weight; is the sample index score.

[0251] Furthermore, the mathematical representation of the weighted aggregation model is as follows:

[0252]

[0253] where is the independent evaluation index of node ; is the dynamic security score of is the comprehensive weight of the independent evaluation index ; is the comprehensive dynamic security score of node ; is the total number of independent evaluation indexes.

[0254] Taking the monthly crude oil shipping network data from January 2014 to December 2023 as an example, the implementation process of the systematic quantitative evaluation method for maritime network security is carried out. Regarding data input and evaluation index calculation, maritime network security level division, key index screening and weight determination, inverse distance dynamic index scoring, and systematic quantitative evaluation of maritime network security, the specific process includes:

[0255] (I) Data input and evaluation index calculation

[0256] Input the continuous monthly crude oil shipping network data from January 2014 to December 2023 and other relevant data on security impact factors. The length of the network time series is , and a total of 82 nodes participate in the network. Calculate the material security indexes of the crude oil shipping network, including: import stability , export stability , trade dependence , material concentration ; topological structure security indexes, including: ratio of the largest connected subgraph , network clustering coefficient , network average degree , network density , modularity ; dynamic evolution security indexes, including: change ratio of the ratio of the largest connected subgraph , change ratio of the network clustering coefficient , change ratio of the network average degree , change ratio of the network density , change ratio of the modularity , node dynamic evolution stability , edge dynamic evolution stability ; geopolitical environment security indexes, including: network trade balance value , earthquake disturbance , sea ice disturbance 、 Sea surface height 、 Consumer price 、 Geopolitical risk 。

[0257] (II) Classification of the security level of the maritime transportation network

[0258] For the time series of the crude oil maritime transportation network, based on the functional safety and structural safety of the network, a two-dimensional classification criterion is established:

[0259] For each network in the time series of the crude oil maritime transportation network, calculate its network resilience according to the network resilience calculation formula; for the crude oil maritime transportation network in December 2023, calculate its network resilience according to the network resilience calculation formula ;

[0260] For each network in the time series of the crude oil maritime transportation network, calculate its monthly total network throughput according to the total network throughput calculation formula; for the crude oil maritime transportation network in December 2023, calculate the total network throughput within one month according to the total network throughput calculation formula ;

[0261] Calculate the network resilience of the crude oil maritime transportation network according to the network resilience calculation formula, and obtain that the average value of the network resilience of the crude oil maritime transportation network is 1.85 and the standard deviation is 0.36. Classify the structural safety level of the crude oil maritime transportation network:

[0262]

[0263] Calculate the monthly total network throughput of the crude oil maritime transportation network according to the total network throughput calculation formula, and obtain the average value of the monthly total throughput of the crude oil maritime transportation network and the standard deviation 。 Classify the functional safety level of the crude oil maritime transportation network:

[0264]

[0265] As Figure 4 shown, Figure 4 is a schematic diagram of the calculation results of the network security level in the embodiment of the present invention. Based on the comprehensive structural safety and functional safety of the crude oil maritime transportation network, classify the overall security level of the maritime transportation network:

[0266]

[0267] For the crude oil maritime transportation network in December 2023, its network resilience , the total network throughput , and its overall security level of the maritime transportation network is the high level.

[0268] (III) Screening of key indicators and determination of weights

[0269] Step 31 Detect the multicollinearity among indicators through the variance inflation factor (VIF), and exclude redundant indicators with high correlations. Each indicator of the crude oil maritime transportation network is detected by VIF, and the network average degree ( ), modularity ( ), network density ( ), export stability ( ) are excluded; and the remaining indicators are processed by Z-score standardization.

[0270] Step 32 Based on the data after removing collinearity, analyze the differences of indicators under different safety levels (low, medium, high) through ANOVA, and screen out the indicators that are statistically significant for safety level classification. The samples are divided into three groups according to the safety level (low, medium, high): low safety ( ), medium safety ( ), high safety ( ); the number of samples in each group is , , , and the total number of samples is ; the ANOVA analysis results and correction results are shown in Table 1.

[0271] Table 1 ANOVA analysis results and correction summary table

[0272]

[0273] Step 33 Use the standardized data set , quantify the contribution of each indicator to the safety level classification through the permutation importance of the random forest model, and screen out the key safety evaluation indicators. The dimension of the feature vector in the model is ; the random forest consists of decision trees; the number of permutation repetitions is ; ; the node splitting criterion is Gini impurity, and the calculation results are shown in Table 2.

[0274] Table 2 Random forest permutation importance

[0275]

[0276] Step 34 By combining ANOVA significance analysis and random forest permutation importance, screen out the key indicators and calculate the comprehensive weights of the indicators , and the calculation results are shown in Table 3.

[0277] Table 3 Indicator comprehensive weights

[0278]

[0279] (4) Anti - distance dynamic index scoring

[0280] Step 41: For each safety level, calculate the standardized mean center of the key indicators, and the calculation results are shown in Table 4.

[0281] Table 4 Standardized mean center of key indicators

[0282]

[0283] Step 42: For each indicator of all crude oil shipping networks, calculate its dynamic safety score according to the distance calculation formula to the level center, the distance weighting formula, and the score calculation formula. For the low level, the benchmark score is 0, for the medium level, the benchmark score is 1, and for the high level, the benchmark score is 2. For the crude oil shipping network in December 2023, its standardized import stability , and its dynamic safety score .

[0284] (5) Systematic quantitative evaluation of shipping network safety

[0285] Based on the dynamic scoring results and indicator weights, calculate the overall safety scores of each crude oil shipping network. For the crude oil shipping network in December 2023, the calculation results are as Figure 5 shown, its overall shipping network safety ; material safety ; geopolitical environment safety ; topological structure safety ; dynamic evolution safety .

[0286] The implementation basis of each embodiment of the present invention is achieved through programmed processing by a device with processor functions. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this actual situation, on the basis of the above - mentioned embodiments, an embodiment of the present invention provides a shipping network security evaluation model based on hybrid weights and anti - distance center aggregation, which is used to execute a shipping network security evaluation method based on hybrid weights and anti - distance center aggregation in the above - mentioned method embodiments.

[0287] See Figure 6 , this model includes:

[0288] Data acquisition and processing module, responsible for collecting network data of the shipping network and safety impact factors, as well as calculating the importance of each node and various network evaluation indicators;

[0289] A security level classification module for classifying the security level of a network;

[0290] A standardized data set construction module that screens independent evaluation indicators and constructs a standard data set;

[0291] A comprehensive weight calculation module for obtaining the comprehensive weights of independent evaluation indicators;

[0292] A dynamic scoring module for obtaining the dynamic security scores of independent evaluation indicators;

[0293] A security score output module for outputting the overall security score of the maritime network.

[0294] It should be noted that the model embodiments provided by the present invention, in addition to being used to implement the methods in the above method embodiments, are also used to implement the methods in other method embodiments provided by the present invention. The difference lies only in setting the corresponding functional modules, and its principle is basically the same as the principle of the above model embodiments provided by the present invention. As long as those skilled in the art, based on the above model embodiments, refer to the specific technical solutions in other method embodiments, obtain the corresponding technical means by combining technical features, and the technical solutions constituted by these technical means, and on the premise of ensuring the practicability of the technical solutions, improve the devices in the above model embodiments to obtain corresponding model class embodiments for implementing the methods in other method class embodiments.

[0295] The model embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, located in one place, or distributed to multiple network units. Select some or all of the modules according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0296] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0297] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in the flow Figure 1 a flow or flows and / or blocks Figure 1 a block or blocks.

[0298] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means for implementing the functions specified in the flow Figure 1 a flow or flows and / or blocks Figure 1 a block or blocks.

[0299] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the flow Figure 1 a flow or flows and / or blocks Figure 1 a block or blocks.

[0300] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features. And these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the security of a maritime transportation network based on hybrid weights and inverse distance center aggregation, characterized in that, Including: Construct a multi-level evaluation system for the maritime transport network, obtain the network data of the maritime transport network and the safety impact factors, calculate the importance of each node in the maritime transport network, and then calculate each network evaluation index in the multi-level evaluation system; Construct a benchmark for classifying the safety level of the network and classify the safety level of the network; Screen out independent evaluation indicators based on the variance inflation factor and perform standardization processing, group them based on the safety level of the network, and then construct a standardized data set; Process the standardized data set using analysis of variance and a random forest model respectively, and the processing results are weighted and fused to obtain the comprehensive weights of each independent evaluation indicator; Based on the standardized data set, use the inverse distance dynamic mapping mechanism to calculate the dynamic safety scores of each independent evaluation indicator; Construct a weighted aggregation model to summarize the dynamic safety scores and mixed weights of each independent indicator, and output the overall safety score of the maritime transport network.

2. The method for evaluating the security of a maritime transportation network based on hybrid weight and inverse distance center aggregation as claimed in claim 1, wherein The multi-level evaluation system includes four first-level safety core evaluation dimensions, including material safety, topological structure safety, dynamic evolution safety, and geopolitical environment safety; Material safety includes four second-level evaluation indicators, namely import stability, export stability, trade dependence, and material concentration; Topological structure safety includes five second-level evaluation indicators, namely modularity, ratio of the largest connected subgraph, network clustering coefficient, network average degree, and network density; Dynamic evolution safety includes seven second-level evaluation indicators, namely node dynamic evolution stability, edge dynamic evolution stability, modularity change ratio, ratio of the largest connected subgraph change ratio, network clustering coefficient change ratio, network average degree change ratio, and network density change ratio; Geopolitical environment safety includes six second-level evaluation indicators, namely network trade balance value, earthquake disturbance, sea ice disturbance, sea surface height, geopolitical risk, and consumer price; 3. The method for evaluating the security of a maritime transportation network based on hybrid weight and inverse distance center aggregation according to claim 2, wherein, Network data includes the historical monthly OD connection quantity and the historical monthly OD trade quantity of the maritime transport; Safety impact factors include, but are not limited to, earthquake disturbance, sea ice disturbance, geopolitical risk, consumer price, and network trade balance value of the nodes in the network; 4. The method for evaluating the security of a maritime transportation network based on hybrid weight and inverse distance center aggregation according to claim 1, wherein, The calculation steps of the node importance are as follows: Based on the topological structure of the network and the historical monthly OD connection quantity of the nodes, calculate the normalized degree centrality, normalized closeness centrality, and normalized betweenness centrality evaluation indicators of each node, sort all the nodes of the maritime transport network in different central point dimensions according to the calculation results, and calculate the scores of each node in different central point dimensions according to the sorting situation; Fuse the scores of each node in different central point dimensions by weighting to calculate the comprehensive centrality score of the node, and normalize the total sum of the comprehensive scores of all nodes in the network as the node importance.

5. The maritime network security evaluation method based on hybrid weight and inverse distance center aggregation as claimed in claim 1, wherein The benchmark for classifying the safety level of the maritime transport network includes two aspects: functional safety and structural safety. Specifically: Establish a network security level classification benchmark for functional safety and structural safety respectively, and divide the maritime transport network into three security levels: 0, 1, and 2; Integrate the structural safety level and functional safety level of the maritime transport network to classify the overall safety level of the maritime transport network. The classification rules are as follows: ； Among them, and represent the structural safety level and functional safety level of the maritime transportation network respectively.

6. The marine network security evaluation method based on hybrid weight and inverse distance center aggregation as claimed in claim 1, wherein, The specific process of screening based on the variance inflation factor is as follows: Construct a linear regression model for each evaluation index of the network, calculate the variance inflation factor, set the maximum threshold of the variance inflation factor for data cleaning, and retain the evaluation indexes as independent evaluation indexes.

7. The method for evaluating the security of a maritime transportation network based on hybrid weight and inverse distance center aggregation according to claim 1, wherein The process of obtaining the comprehensive weight of the evaluation indexes of the independent evaluation indexes is specifically as follows: For the standardized data set, calculate the within-group variance and between-group variance of different network security level groups for each independent evaluation index respectively, then calculate the F statistic and p value using the within-group variance and between-group variance, and use multiple test correction of the p value to avoid false positives. Finally, select the independent evaluation indexes with statistical significance and output the corrected p value of the indexes. Input the standardized data set to train the random forest model. The trained model outputs the feature importance score of the independent evaluation indexes based on the Gini importance of the features for model credibility verification, and at the same time outputs the permutation importance score of the independent evaluation indexes. Weightedly fuse the corrected p value of the independent evaluation indexes with statistical significance and the permutation importance score of the independent evaluation indexes to obtain the comprehensive weight of the independent evaluation indexes.

8. A method for evaluating the security of a maritime transportation network based on hybrid weight and inverse distance center aggregation as described in claim 1, characterized in that, The steps of calculating the dynamic security score of each independent evaluation index using the inverse distance dynamic mapping mechanism are specifically as follows: For the standardized data set, for different network security level groups, calculate the standardized mean center of each evaluation index within the group respectively. Calculate the Euclidean distance from the evaluation index within the group to the standardized mean center according to the standardized mean center, and calculate the normalized weight of the evaluation index for the network security level in combination with the smoothing coefficient. Construct an inverse distance weight dynamic scoring model, and output the dynamic security score of the evaluation index in combination with the normalized weight of the evaluation index at different security levels.

9. The security evaluation method for a maritime transportation network based on hybrid weight and inverse distance center aggregation as claimed in claim 1, wherein, The mathematical representation of the weighted aggregation model is as follows: ； Among them, is the independent evaluation index of the network of the dynamic security score; is the comprehensive weight of the independent evaluation index ; is the comprehensive dynamic security score of the network ; is the total number of independent evaluation indicators.

10. A maritime network security evaluation model based on hybrid weights and inverse distance center aggregation, characterized in that, A maritime network security evaluation method based on the aggregation of mixed weights and inverse distance centers, which is used to implement any one of claims 1-9, includes: A data collection and processing module, which is responsible for collecting the network data of the maritime network and the security impact factors, and calculating the importance of each node and various network evaluation indexes. A security level division module, which is used to divide the security level of the network. A standardized data set construction module, which screens independent evaluation indexes and constructs a standard data set. A comprehensive weight calculation module, which obtains the comprehensive weight of the independent evaluation indexes. A dynamic scoring module, which obtains the dynamic security score of the independent evaluation indexes. A security score output module, which outputs the overall security score of the maritime network.

Images