Key node risk identification method and system for industrial chain

Abstract

The invention discloses a key node risk identification method and system for an industrial chain. The method comprises the following steps: acquiring industrial chain node information, establishing an industrial chain network according to an upstream and downstream relationship of the industrial chain nodes, and determining industry and enterprise information of each industrial chain node; determining a static correlation degree between industrial chain nodes according to a whole-industry input-output table, and determining a dynamic correlation degree between the industrial chain nodes according to the power consumption data; according to the static correlation degree and the dynamic correlation degree, constructing node characteristics, power utilization characteristics and toughness characteristics of industrial chain nodes, and determining key nodes in the industrial chain network; and performing risk monitoring on the identified industrial chain key node, including historical electricity consumption monitoring and enterprise production hysteresis monitoring, and identifying a risk node based on a risk monitoring result. According to the scheme, the accuracy of risk monitoring is improved.

Description

Technical Field

[0001] The present invention belongs to the field of industrial chain risk point identification, and in particular relates to a method and system for identifying key node risks in an industrial chain. Background Art

[0002] With the development of a new round of industrial transformation, the global industrial chain and value chain are facing new challenges. By identifying the risk points in the industrial chain, we can better control the overall development of the industrial chain, effectively avoid the risks of the industrial chain, and enhance the resilience and security of the industrial chain.

[0003] At present, the risk assessment of the industrial chain mainly focuses on qualitative and quantitative aspects. Among them, the qualitative method mainly analyzes based on expert experience, establishes an evaluation system and detailed indicators, and finally conducts a comprehensive assessment based on expert scoring. It requires a lot of manpower and material resources, and because it relies on expert experience and professional knowledge, it is easy to cause deviations in the assessment results. The quantitative method is more widely used because it is more objective and accurate. In the prior art, CN117610909A proposes a multi-cycle risk assessment method and system for the industry chain based on big data, obtains the current information data of the enterprise and the information data of the newly opened project, extracts multiple industry information data from the industry information database according to the industry type, the government policy of the location of the newly opened project, and the external environment, and first conducts a separate risk analysis and assessment for each acquired information data, and then conducts a comprehensive assessment. CN117575328A proposes an industrial chain risk assessment method and device based on a graph attention neural network, which aims to solve the problem of poor accuracy in the existing industrial chain risk assessment method, and constructs the network topology structure of the upstream, midstream and downstream of the industrial chain according to the data information of the industrial chain to be evaluated; constructs a graph attention neural network assessment model and performs model training, and uses the trained graph attention network assessment model to calculate the node risk value and attention relationship coefficient of the industrial chain network to be evaluated. CN117634889A proposes an industrial chain risk warning assessment method based on PCA-CNN. On the basis of establishing an industrial chain financial risk warning indicator system, it combines the convolutional neural network method to construct an industrial chain risk warning model, and uses the principal component analysis method to measure the evaluation indicators. The relevant data of listed companies in the industrial chain are selected to conduct an empirical analysis of the model.

[0004] In general, the traditional quantitative identification of industrial chain risks mainly focuses on the financial data and basic information of enterprises. However, these data are updated slowly, and there is a lag in risk identification. Secondly, traditional modeling methods are to identify risks for all enterprises in the industrial chain. For the industrial chain, there are key nodes in each link, and occasional risks at non-key nodes have little impact on the entire industrial chain. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides a method and system for identifying risks of key nodes in the industrial chain, which constructs a static relationship of the input-output table and a dynamic correlation relationship based on electricity consumption based on the existing industrial chain network, and measures the correlation of the industrial chain nodes by combining dynamic and static indicators; extracts key nodes based on the industrial chain network characteristics and electricity consumption characteristics, and conducts risk monitoring on them.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions.

[0007] The present invention first discloses a method for identifying risks at key nodes of an industrial chain, the method comprising the following steps:

[0008] Step 1: Obtain information on industry chain nodes based on predefined classification rules, establish an industry chain network based on the upstream and downstream relationships of industry chain nodes, and determine the industry and enterprise information of each industry chain node based on the patent database and industrial and commercial information database;

[0009] Step 2: Based on the upstream and downstream relationship of the industrial chain nodes, the static correlation between the industrial chain nodes is determined according to the input-output table of the entire industry, and the dynamic correlation between the industrial chain nodes is determined according to the electricity consumption data;

[0010] Step 3: construct node characteristics, power consumption characteristics, and resilience characteristics of the nodes in the industrial chain according to the static correlation and dynamic correlation between the nodes in the industrial chain, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics, and resilience characteristics;

[0011] Step 4: Perform risk monitoring on the identified key nodes of the industrial chain, wherein the risk monitoring includes historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

[0012] The present invention further includes the following preferred embodiments:

[0013] Determining the static correlation between the nodes of the industrial chain according to the input-output table of the entire industry further includes:

[0014] Combined with the N industries involved in the industrial chain network, the corresponding industry names are searched in the input-output table of the entire industry to obtain the static correlation matrix B between industries;

[0015] Performing standardization on the static association relationship matrix B;

[0016]

[0017] x is the original data in matrix B, x min is the minimum value of all data in matrix B, x maxis the maximum value in matrix B;

[0018] Normalize each row in the matrix B to obtain the normalized static correlation matrix B new .

[0019] Determining the dynamic correlation between the nodes in the industrial chain according to the electricity consumption data further includes:

[0020] Obtain the monthly electricity consumption of each industry or enterprise in the past two years; use the Granger causality analysis algorithm, assume that the time series of electricity consumption data of any two industries or enterprises are Y and X, and establish the following autoregressive model with lag terms:

[0021]

[0022] X t is the vector formed by the electricity consumption of the electricity consumption data sequence X in period t, Y t is the vector of electricity consumption of enterprise Y in period t, α 10 and α 20 is the intercept term, α 11 (q) , α 21 (q) , α 12 (p) , α 22 (p) are regression coefficients of various orders, where p, q = 1, 2, 3...n, n is the number of lag periods, ε 1t and ε 2t is the error term, and the original hypothesis H0:α 12 (1) =α 12 (2) =... =α 12 (p) =0, that is, X t Not Y t The reason for this is to conduct an F test on the regression coefficient. The F statistic is as follows:

[0023]

[0024] Where F follows the degrees of freedom (q, npq-1), SSR1 is the time series Y t The residual sum of squares in the constructed autoregressive model, SSR2 is the residual sum of squares after adding the lag term X t The residual sum of squares of the binary autoregressive model after ;

[0025] At the significance level of 0.05, the P value was calculated based on the F test, and its value was determined to be less than 0.05;

[0026] If the P value is less than 0.05, the null hypothesis is rejected, indicating that X t Yes t The reason is that the test statistic F value is used as the weight measure of the edge between nodes;

[0027] If the P value is greater than 0.05, the null hypothesis is accepted, indicating that X t Not Y t The reason is that the weight value between nodes is 0;

[0028] The above regression analysis is performed on the electricity consumption sequences of N industries or enterprises in pairs, and the weight values ​​between any two industries or enterprises are obtained to form a dynamic correlation matrix E;

[0029] The original static correlation matrix is ​​modified using the information of the dynamic correlation matrix E to obtain the correlation between each node in the industrial chain network:

[0030] F=B new ·E

[0031] Normalize the corrected correlation F again to get the final correlation value F new .

[0032] The node features of the industrial chain network are further constructed as follows:

[0033] Calculate the betweenness centrality B of the industrial chain node:

[0034]

[0035] Among them, n kj For node a k to a j The number of shortest paths between kj (i) Represented as node a k to a j Passing through node a i The number of shortest paths;

[0036] Calculate the proximity centrality index C of the industrial chain node:

[0037]

[0038] in, Represents node a i The sum of the shortest distances to other nodes;

[0039] Calculate the degree centrality D of the industrial chain node:

[0040]

[0041] Among them, N-1 represents the maximum possible degree of the industrial chain network, d i Represents node a i degree of node a i The sum of the number of all production nodes that are connected is node a i The degree of i ;

[0042] The betweenness centrality B, proximity centrality index C and degree centrality D are taken as the node characteristics of the industrial chain network.

[0043] The construction of the electricity consumption characteristics of the industrial chain network further includes:

[0044] Calculate the electricity consumption proportion E of the industrial chain network:

[0045]

[0046] Where P i represents the annual electricity consumption of enterprise i, It represents the sum of annual electricity consumption of all N enterprises in the industry to which the enterprise belongs;

[0047] Standardize electricity usage data:

[0048]

[0049] x is the original power consumption data, x min is the minimum value of all data in the power consumption curve, x max is the maximum value in the power consumption curve, x new is the value of x after normalization;

[0050] Calculate the stability characteristics of electricity consumption data W of the industrial chain network:

[0051] W;

[0052] The electricity consumption proportion E and the electricity consumption data stability characteristic F are used as the electricity consumption characteristics of the industrial chain network.

[0053] In step 3, the constructing of the resilience characteristics of the industrial chain network further includes:

[0054] Calculate the global network connectivity efficiency P of the industrial chain network:

[0055]

[0056] Among them, d ij For any two industrial chain nodes v i to v j The correlation value between them, N is the number of industries or enterprises in the industrial chain network;

[0057] Calculate the maximum connected subgraph size S of the industrial chain network, that is, the ratio of the number of nodes contained in the largest connected subgraph to the total number of nodes in the network:

[0058]

[0059] Among them, L represents the number of nodes contained in the largest connected subgraph; N is the number of industries or enterprises in the industrial chain network, that is, the number of nodes.

[0060] The global network connectivity efficiency P and the maximum connected subgraph size S are used as the resilience characteristics of the industrial chain network.

[0061] The step 4 further comprises:

[0062] Step 4.1, obtain the enterprise's historical electricity consumption data up to the latest month, and perform z-score standardization on the historical monthly electricity consumption data:

[0063]

[0064] Among them, μ is the mean of the series, σ is the standard deviation of the series, and x is the electricity consumption data;

[0065] If the absolute value of the standardized current monthly data is greater than 3, it is determined that the enterprise has production abnormalities;

[0066] Step 4.2, obtain the monthly electricity consumption of the industry to which the enterprise belongs in a specific period H = [h1, h2...h q ];

[0067] Get the electricity consumption of the enterprise within the cycle and the previous 12 periods, that is,

[0068] Y=[y -12 ,y -11 ...y1,y2,..y q ];

[0069] Calculate the correlation between two vectors shifted by 12 periods in a cycle; perform sliding translation 12 times, each time moving forward one period, to obtain a new Y sequence. Calculate the correlation between the 12 sequences and the H sequence respectively. The translation period corresponding to the maximum correlation value is the lag period. If the lag period is greater than the preset threshold, it is determined that there is a risk anomaly.

[0070] Step 4.3: If any of steps 4.1 and 4.2 identifies that the enterprise has risk anomalies, then the enterprise is determined to have risk anomalies.

[0071] The present invention also discloses a key node risk identification system for an industrial chain using the aforementioned key node risk identification method for an industrial chain, comprising:

[0072] The industrial chain network construction module is used to obtain industrial chain node information based on predefined classification rules, establish an industrial chain network according to the upstream and downstream relationships of industrial chain nodes, and determine the industry and enterprise information of each industrial chain node based on the patent database and industrial and commercial information database;

[0073] A correlation calculation module is used to determine the static correlation between the nodes in the industrial chain based on the upstream and downstream relationships of the nodes in the industrial chain and the input-output table of the entire industry, and to determine the dynamic correlation between the nodes in the industrial chain based on the power consumption data;

[0074] A key node identification module, used to construct node characteristics, power consumption characteristics and resilience characteristics of the industrial chain nodes according to the static correlation and dynamic correlation between the industrial chain nodes, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics and resilience characteristics;

[0075] The risk identification module is used to perform risk monitoring on the identified key nodes of the industrial chain, wherein the risk monitoring includes historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

[0076] Accordingly, the present application also discloses a terminal, including a processor and a storage medium;

[0077] The storage medium is used to store instructions;

[0078] The processor is used to operate according to the instructions to execute the steps of the aforementioned method for identifying risks of key nodes in the industrial chain.

[0079] Correspondingly, the present application also discloses a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the aforementioned method for identifying risks at key nodes in the industrial chain are implemented.

[0080] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a method and system for identifying risks of key nodes in an industrial chain, extracts key nodes based on the network characteristics and electricity consumption characteristics of the industrial chain, realizes risk monitoring of the industrial chain, and improves the accuracy of risk monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 It is a flow chart of the industrial chain risk identification method in the present invention.

[0082] Figure 2 It is the importance index of the industrial chain network node in the present invention.

[0083] Figure 3 It is the industrial chain network hierarchy structure in the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solution and advantages of the present invention more clear, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0085] The embodiments described in this application are only some embodiments of the present invention, not all embodiments. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the protection scope of the present invention.

[0086] In view of the shortcomings of the existing technology, the present invention proposes a method and system for identifying risks of key nodes in the industrial chain, which realizes the construction of an industrial chain based on electricity consumption, the extraction of key nodes based on network characteristics and electricity consumption characteristics, and the risk monitoring and identification of key characteristics.

[0087] See also Figure 1 As shown, the method for identifying risks at key nodes of an industrial chain disclosed in the present invention comprises the following steps:

[0088] Step 1: Obtain the information of industrial chain nodes based on predefined classification rules, establish an industrial chain network according to the upstream and downstream relationships of the industrial chain nodes, and determine the industry and enterprise information of each industrial chain node based on the patent database and industrial and commercial information database.

[0089] The theoretical method of industrial chain construction is completed by decomposing the product life cycle and the importance of the correlation between industrial nodes, with reference to the classification rules of domestic key industrial research institutions such as "Foresight Industry Research Institute" and "New Materials Online".

[0090] The sources of industrial chain enterprises mainly include two parts. First, based on the patent data of the State Intellectual Property Office, the ownership of enterprises is determined by the industrial chain node products involved in the patent layout. Secondly, using the industrial and commercial information of the National Enterprise Credit Information Publicity System, mainly processing the enterprises that have opened businesses and include industrial chain node products in their business scope.

[0091] Step 2: Based on the upstream and downstream relationships of the industrial chain nodes, determine the static correlation between the industrial chain nodes according to the industry input-output table, and determine the dynamic correlation between the industrial chain nodes according to the electricity consumption data.

[0092] According to the industrial chain network relationship formed in step 1, the correlation of each node is established, including static and dynamic node correlation. The static relationship is mainly based on the industry input-output table A published by the National Bureau of Statistics. ij, from the column direction, it indicates the consumption of product i in the production of product j, and from the row direction, it indicates the allocation of product i to the production of product j. This part reflects the input-output relationship between product departments and directly reflects the interdependence between various industries. The dynamic association is mainly reflected by the correlation of electricity consumption data of various industries and enterprises. The electricity consumption data represents the changes in the production capacity of the industry and enterprise to a certain extent. The change relationship between electricity consumption between upstream and downstream reflects the correlation of upstream and downstream production capacity.

[0093] Preferably, step 2 comprises the following steps:

[0094] Step 2.1, combining the N industries to which the enterprises involved in the industrial chain network belong, find the corresponding industry name in the whole society input-output table, and obtain the static relationship matrix B between industries;

[0095] Step 2.2: In order to further combine with the dynamic relationship, the static relationship matrix B obtained in step 2.1 is normalized. The processing method includes but is not limited to normalizing each row in the matrix B to obtain the normalized static relationship matrix B. new :

[0096]

[0097] Among them, x is the original data in matrix B, x min is the minimum value of all data in matrix B, x max is the maximum value in matrix B, x new is the value of x after normalization.

[0098] Step 2.3, obtain the monthly electricity consumption of various industries and enterprises involved in the industrial chain network in the past two years. Due to the dynamic fluctuations of the influence of upstream and downstream enterprises and the lag of the market, directly measuring the correlation of the electricity consumption curves of the two industries cannot measure the real correlation between the industries well. Granger causality analysis is used in the present invention to incorporate lag into the autoregressive model with lag terms.

[0099] Specifically, the electricity consumption data series of the two industries and enterprises are Y and X respectively. The following autoregressive model with lag term is established for the time series Y and X:

[0100]

[0101] X t is the vector formed by the electricity consumption of the electricity consumption data sequence X in period t, Y t is the vector of electricity consumption of enterprise Y in period t, and is α 10 and α 20 is the intercept term, α 11 (q), α 21 (q) , α 12 (p) , α 22 (p) are regression coefficients of various orders, where p, q = 1, 2, 3...n, n is the number of lag periods, ε 1t and ε 2t is the error term, and the original hypothesis H0:α 12 (1) =α 12 (2) =... =α 12 (p) =0, that is, X t Not Y t The reason for this is to conduct an F test on the regression coefficient. The F statistic is as follows:

[0102]

[0103] Where F follows the degrees of freedom (q, npq-1), SSR1 is the time series Y t The residual sum of squares in the constructed autoregressive model, SSR2 is the residual sum of squares after adding the lag term X t The residual sum of squares of the binary autoregressive model after . At the significance level of 0.05, the P value is calculated based on the F test, and its value is determined to be less than 0.05.

[0104] If the P value is less than 0.05, the null hypothesis is rejected. t Yes t The reason is that the test statistic F value is used as the weight measure of the edge between nodes;

[0105] If the P value is greater than 0.05, the null hypothesis is accepted. t Not Y t The reason is that the weight value between the nodes is 0.

[0106] The electricity consumption series of N industries and enterprises are regressed pairwise to obtain the weight values ​​between them, forming a dynamic correlation matrix E.

[0107] Step 2.4, using the static correlation matrix and dynamic correlation matrix between industries obtained above, the original static correlation matrix is ​​corrected through the dynamic correlation matrix information to obtain the correlation between each node in the industrial chain network.

[0108] F=B new ·E

[0109] Normalize the corrected correlation F again to get the final correlation value F new .

[0110] Step 3: According to the static correlation and dynamic correlation between the nodes in the industrial chain, construct the node characteristics, power consumption characteristics and resilience characteristics of the industrial chain nodes, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics and resilience characteristics.

[0111] Based on the industrial chain network and node correlation information obtained in steps 1 and 2, key indicators of each node in the industrial chain are established, including three dimensions, namely node characteristics, power consumption characteristics, and resilience characteristics, such as Figure 2 shown.

[0112] Preferably, step 3 comprises the following specific steps:

[0113] Step 3.1, first construct node features. The industrial chain network is abstracted into a network G(V, E) consisting of a set of production nodes V (industries or enterprises) and a set of edges E (the degree of association between nodes). The following properties of the network can characterize the performance of industrial chain nodes:

[0114] 1) Betweenness centrality B. This index indicates the ability of a specific node in the network to control other nodes. The larger the value, the stronger the transit function of the node, and the more obvious its role as a bridge in the entire network. The formula is as follows:

[0115]

[0116] Among them, n kj For node a k to a j The number of shortest paths between kj (i) Represented as node a k to a j Passing through node a i is the number of shortest paths, and N is the total number of nodes in the network graph.

[0117] 2) Proximity centrality index C. This index indicates the degree to which a node is central in the network. The greater the proximity centrality, the stronger the closeness between nodes. The formula is as follows:

[0118]

[0119] in, Represents node a i The sum of the shortest distances to other nodes, where N is the total number of nodes in the network graph.

[0120] 3) Degree centrality D. This index represents the centrality of the node. The larger the degree centrality of the node, the higher the degree centrality of the node, and the more important the node is in the network. i The sum of the number of all production nodes that are connected is node a iThe degree of i , the formula is as follows:

[0121]

[0122] Among them, N-1 represents the maximum possible degree of the industrial chain network, d i Represents node a i The degree.

[0123] Step 3.2, then construct the electricity consumption characteristics of each node in the industrial chain network. Through the enterprise electricity consumption data, calculate the importance of the enterprise in the industry, including the following indicators:

[0124] 1) Electricity consumption proportion E. For enterprises in the same industry, their electricity consumption can measure the production capacity of the enterprise to a certain extent. The proportion of the enterprise's electricity consumption in the industry can measure the importance of the enterprise in the entire industry. The larger the proportion of the enterprise, the greater the impact of changes in its production capacity on the industrial chain.

[0125]

[0126] Where P i represents the annual electricity consumption of enterprise i, It represents the sum of annual electricity consumption of all N enterprises in the industry to which the enterprise belongs.

[0127] 2) Power consumption stability F indicates that the enterprise can stably provide corresponding products to the industrial chain, which is particularly important to the industrial chain. The variance of the enterprise's power consumption curve can measure its fluctuation. Under the same time dimension, the larger the variance, the greater the data fluctuation and the more unstable it is. In the future, to avoid the impact of power consumption on stability, the power consumption data should be standardized before stability calculation.

[0128]

[0129] x is the original power consumption data, x min is the minimum value of all data in the power consumption curve, x max is the maximum value in the power consumption curve, x new is the value of x after normalization.

[0130]

[0131] Among them, x i is the electricity consumption of the enterprise in the i-th month after standardization, It is the average standardized electricity consumption of an enterprise in a time period, and T is the number of months in a period.

[0132] Step 3.3, determine the resilience characteristics of the industrial chain from the operating status of the industrial chain network itself. The impact of the existence or absence of a node on the industrial chain measures the importance of the node.

[0133] 1) Global network connectivity efficiency P. This connectivity efficiency value integrates the connection efficiency between all nodes in the network with paths:

[0134]

[0135] Among them, d ij For any two industrial chain nodes v i to v j The correlation value between new It is obtained from, N is the number of industries or enterprises in the industrial chain network.

[0136] 2) Maximum connected subgraph size S. It is defined as the ratio of the number of nodes contained in the largest connected subgraph to the total number of nodes in the network. As the network changes under the influence of different policies, this attribute can reflect the stability of the network.

[0137]

[0138] Among them, L represents the number of nodes contained in the largest connected subgraph. N is the number of industries or enterprises in the industrial chain network, that is, the number of nodes.

[0139] The indicators P and S measure the stability of the industrial chain network, based on which the node importance indicators ΔP and ΔS are constructed.

[0140]

[0141] Where ΔP i and ΔS i They represent the changes in the connectivity efficiency and the size of the largest connected subgraph of the industrial chain network before and after the removal of the node enterprise i, and measure the importance of the enterprise to the industrial chain. 前 and S 前 They represent the global network connectivity efficiency P of the industrial chain before removing the corresponding nodes and the maximum connected subgraph size S, P 后 and S 后 They respectively represent the global network connectivity efficiency P and the maximum connected subgraph size S of the industrial chain before removing the corresponding nodes.

[0142] Step 3.4, finally select the key nodes in the industrial chain from the above three types of indicators for measuring the importance of enterprises. Considering the need to identify the key nodes at different levels of the industrial chain, the identification process is carried out layer by layer according to the industrial chain level. Specifically, it includes the following steps:

[0143] 1) For all enterprise nodes in the industrial chain, calculate their corresponding importance eigenvalue matrix I, I = [B, C, D, E, F, ΔP, ΔS].

[0144] 2) For each industry node, such as Figure 3 As shown in the figure, there are m×n×p industry nodes in total. The importance values ​​of their subordinate enterprise nodes are compared. The importance values ​​are obtained according to the importance feature matrix and are calculated as follows:

[0145] Importance=(E>10%)×(B+C+D+F+ΔP+ΔS)

[0146] The E index in the importance feature measures the size of the enterprise, and the enterprise index measures the node's own attributes, including its correlation with others, its own stability, and its contribution to the overall industrial chain. However, all these attributes are linked to its size. For a micro-sized enterprise, even if it has a close relationship with each node, the risk impact it ultimately generates is relatively small for the entire industrial chain. Therefore, enterprises of a certain size are initially screened out, and then the sum of the important attributes of the enterprise is calculated.

[0147] 3) For each industry, calculate the importance of its subordinate enterprises respectively, and sort and select the top T enterprises as key nodes. The T value is specified based on experience. If more enterprise node risks need to be paid attention to, the T value is set larger, otherwise it is set smaller.

[0148] Step 4: Conduct risk monitoring on the identified key nodes of the industrial chain, including historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

[0149] For the key nodes selected in step 3, risk monitoring is carried out by monitoring the changes in electricity consumption. Using electricity consumption indicators as monitoring has the advantage of timeliness. Compared with using financial data, it can detect problems and control risks more timely.

[0150] Preferably, step 4 specifically includes the following steps:

[0151] Step 4.1, identify the enterprise's own risks by monitoring the enterprise's historical electricity consumption and identifying abnormal electricity consumption. The outlier identification method can be based on the 3σ criterion, but is not limited to this method. Other methods such as clustering and box plots can also be used to carry out abnormal electricity consumption analysis.

[0152] Obtain the enterprise's historical electricity consumption data up to the latest month, and perform z-score standardization on the historical monthly electricity consumption data.

[0153]

[0154] Among them, μ is the mean of the series, σ is the standard deviation of the series, and x is the electricity consumption data.

[0155] If the absolute value of the standardized current monthly data is greater than 3, it is determined that the enterprise has production abnormalities.

[0156] Step 4.2, identify the risks of the enterprise in the industry. Each industry in the industrial chain has its own market development rules. The delay and lag of one link will have an impact on the entire industrial chain. By analyzing the difference between the fluctuation of the enterprise's production capacity and the fluctuation of the overall production capacity of its industry, the production lag of the enterprise is obtained as the risk value.

[0157] 1) Obtain the monthly electricity consumption of the industry to which the enterprise belongs in a specific period H = [h1, h2...h q ].

[0158] 2) Obtain the electricity consumption of the enterprise within the cycle and the previous 12 periods, that is,

[0159] Y=[y -12 ,y -11 ...y1,y2,..y q ].

[0160] 3) Calculate the correlation between two vectors shifted by 12 periods in a loop. The correlation index is not limited to the correlation measurement methods such as cosine distance, Euclidean distance, and Pearson coefficient, among which the Euclidean distance is:

[0161]

[0162] Slide and translate 12 times, each time moving forward one period, to obtain a new Y sequence. Calculate the correlation between the 12 sequences and the H sequence respectively. The translation period corresponding to the maximum correlation value is its lag period. The larger the lag period, the greater the risk value. If the lag period is greater than the preset threshold, it is determined that there is a risk anomaly. The lag period is 0, which means that the production behavior of the enterprise is consistent with the trend of the industry.

[0163] Step 4.3: If any of steps 4.1 and 4.2 identify that the enterprise has risk anomalies, then the enterprise is determined to have risk anomalies and requires further attention.

[0164] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a method and system for identifying risks of key nodes in an industrial chain, extracts key nodes based on the network characteristics and electricity consumption characteristics of the industrial chain, realizes risk monitoring of the industrial chain, and improves the accuracy of risk monitoring.

[0165] The present invention may be a system, a method and / or a computer program product. The present invention also discloses a key node risk identification system for an industrial chain based on the aforementioned key node risk identification method for an industrial chain, comprising:

[0166] The industrial chain network construction module is used to obtain industrial chain node information based on predefined classification rules, establish an industrial chain network according to the upstream and downstream relationships of industrial chain nodes, and determine the industry and enterprise information of each industrial chain node based on the patent database and industrial and commercial information database;

[0167] A correlation calculation module is used to determine the static correlation between the nodes in the industrial chain based on the upstream and downstream relationships of the nodes in the industrial chain and the input-output table of the entire industry, and to determine the dynamic correlation between the nodes in the industrial chain based on the power consumption data;

[0168] A key node identification module, used to construct node characteristics, power consumption characteristics and resilience characteristics of the industrial chain nodes according to the static correlation and dynamic correlation between the industrial chain nodes, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics and resilience characteristics;

[0169] The risk identification module is used to perform risk monitoring on the identified key nodes of the industrial chain, wherein the risk monitoring includes historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

[0170] Based on the spirit of the present invention, those skilled in the art can easily think that a computer program product can be obtained based on the aforementioned risk identification method for key nodes of the industrial chain. The computer program product may include a computer-readable storage medium, which carries computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure. That is, the present application also includes a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the aforementioned risk identification method for key nodes of the industrial chain.

[0171] Computer readable storage medium can be a tangible device that can keep and store the instructions used by the instruction execution device. Computer readable storage medium can be, for example, - but not limited to - electrical storage device, magnetic storage device, optical storage device, electromagnetic storage device, semiconductor storage device or any suitable combination of the above. More specific examples (non-exhaustive list) of computer readable storage medium include: portable computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, for example, punch card or groove protrusion structure with instructions stored thereon and any suitable combination of the above. Computer readable storage medium used here is not interpreted as instantaneous signal itself, such as radio wave or other free propagating electromagnetic wave, electromagnetic wave propagated by waveguide or other transmission medium (for example, light pulse by optical fiber cable) or electrical signal transmitted by wire.

[0172] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0173] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages-such as Smalltalk, C++, etc., and conventional procedural programming languages-such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network-including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA) may be personalized by utilizing the state information of a computer-readable program instruction, and the electronic circuit may execute a computer-readable program instruction, thereby realizing various aspects of the present disclosure.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for identifying risks at key nodes of an industrial chain, characterized in that: The following steps are involved: Step 1: Obtain information on industry chain nodes based on predefined classification rules, establish an industry chain network based on the upstream and downstream relationships of industry chain nodes, and determine the industry and enterprise information of each industry chain node based on the patent database and industrial and commercial information database; Step 2: Based on the upstream and downstream relationship of the industrial chain nodes, the static correlation between the industrial chain nodes is determined according to the input-output table of the entire industry, and the dynamic correlation between the industrial chain nodes is determined according to the electricity consumption data; Step 3: construct node characteristics, power consumption characteristics, and resilience characteristics of the nodes in the industrial chain according to the static correlation and dynamic correlation between the nodes in the industrial chain, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics, and resilience characteristics; Step 4: Perform risk monitoring on the identified key nodes of the industrial chain, wherein the risk monitoring includes historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

2. The multidimensional data analysis method based on industrial chain development according to claim 1 is characterized in that: Determining the static correlation between the nodes of the industrial chain according to the input-output table of the entire industry further includes: Combined with the N industries involved in the industrial chain network, the corresponding industry names are searched in the input-output table of the entire industry to obtain the static correlation matrix B between industries; Performing standardization on the static association relationship matrix B; x is the original data in matrix B, x min is the minimum value of all data in matrix B, x max is the maximum value in matrix B; Normalize each row in the matrix B to obtain the normalized static correlation matrix B new .

3. The multidimensional data analysis method based on industrial chain development according to claim 2 is characterized in that: Determining the dynamic correlation between the nodes in the industrial chain according to the electricity consumption data further includes: Obtain the monthly electricity consumption of each industry or enterprise in the past two years; use the Granger causality analysis algorithm, assume that the time series of electricity consumption data of any two industries or enterprises are Y and X, and establish the following autoregressive model with lag terms: X t is the vector formed by the electricity consumption of the electricity consumption data sequence X in period t, Y t is the vector of electricity consumption of enterprise Y in period t, α 10 and α 20 is the intercept term, α 11 (q) , α 21 (q) , α 12 (p) , α 22 (p) are regression coefficients of various orders, where p, q = 1, 2, 3...n, n is the number of lag periods, ε 1t and ε 2t is the error term, and the original hypothesis H0:α 12 (1) =α 12 (2) =... =α 12 (p) =0, that is, X t Not Y t The reason for this is to conduct an F test on the regression coefficient. The F statistic is as follows: Where F follows the degrees of freedom (q, npq-1), SSR1 is the time series Y t The residual sum of squares in the constructed autoregressive model, SSR2 is the residual sum of squares after adding the lag term X t The residual sum of squares of the binary autoregressive model after ; At the significance level of 0.05, the P value was calculated based on the F test, and its value was determined to be less than 0.05; If the P value is less than 0.05, the null hypothesis is rejected, indicating that X t Yes t The reason is that the test statistic F value is used as the weight measure of the edge between nodes; If the P value is greater than 0.05, the null hypothesis is accepted, indicating that X t Not Y t The reason is that the weight value between nodes is 0; The above regression analysis is performed on the electricity consumption sequences of N industries or enterprises in pairs, and the weight values ​​between any two industries or enterprises are obtained to form a dynamic correlation matrix E; The original static correlation matrix is ​​modified using the information of the dynamic correlation matrix E to obtain the correlation between each node in the industrial chain network: F=B new ·E Normalize the corrected correlation F again to get the final correlation value F new .

4. The method for identifying risks at key nodes of an industrial chain according to claim 3 is characterized in that: In step 3, the node features of the industrial chain network are constructed, further comprising: Calculate the betweenness centrality B of the industrial chain node: Among them, n kj For node a k to a j The number of shortest paths between kj (i) Represented as node a k to a j Passing through node a i The number of shortest paths; N is the total number of nodes in the network graph. Calculate the proximity centrality index C of the industrial chain node: in, Represents node a i The sum of the shortest distances to other nodes; N is the total number of nodes in the network graph. Calculate the degree centrality D of the industrial chain node: Among them, N-1 represents the maximum possible degree of the industrial chain network, d i Represents node a i degree of node a i The sum of the number of all production nodes that are connected is node a i The degree d i ; The betweenness centrality B, proximity centrality index C and degree centrality D are taken as the node characteristics of the industrial chain network.

5. The method for identifying risks at key nodes of an industrial chain according to claim 4 is characterized in that: In step 3, the construction of the electricity consumption characteristics of the industrial chain network further includes: Calculate the electricity consumption proportion E of the industrial chain network: Where P i represents the annual electricity consumption of enterprise i, It represents the sum of annual electricity consumption of all N enterprises in the industry to which the enterprise belongs; Standardize electricity usage data: x is the original power consumption data, x min is the minimum value of all data in the power consumption curve, x max is the maximum value in the power consumption curve, x new is the value of x after normalization; Calculate the stability characteristics of electricity consumption data WF of the industrial chain network: Among them, x i is the electricity consumption of the enterprise in the i-th month after standardization, It is the average value of standardized electricity consumption of an enterprise in a time period, and T is the number of months in a period; The electricity consumption proportion E and the electricity consumption data stability characteristic F are used as the electricity consumption characteristics of the industrial chain network.

6. The multidimensional data analysis method based on industrial chain development according to claim 5 is characterized in that: In step 3, the constructing of the resilience characteristics of the industrial chain network further includes: Calculate the global network connectivity efficiency P of the industrial chain network: Among them, d ij For any two industrial chain nodes v i to v j The correlation value between them, N is the number of industries or enterprises in the industrial chain network; Calculate the maximum connected subgraph size S of the industrial chain network, that is, the ratio of the number of nodes contained in the largest connected subgraph to the total number of nodes in the network: Among them, L represents the number of nodes contained in the largest connected subgraph; N is the number of industries or enterprises in the industrial chain network, that is, the number of nodes. The global network connectivity efficiency P and the maximum connected subgraph size S are used as the resilience characteristics of the industrial chain network.

7. The method for identifying risks at key nodes of an industrial chain according to claim 6 is characterized in that: The step 4 further comprises: Step 4.1, obtain the enterprise's historical electricity consumption data up to the latest month, and perform z-score standardization on the historical monthly electricity consumption data: Among them, μ is the mean of the series, σ is the standard deviation of the series, and x is the electricity consumption data; If the absolute value of the standardized current monthly data is greater than 3, it is determined that the enterprise has production abnormality; Step 4.2, obtain the monthly electricity consumption of the industry to which the enterprise belongs in a specific period H = [h1, h2...h q ]; Get the electricity consumption of the enterprise within the period and the previous 12 periods, that is, Y = [y -12 ,y -11 ...y1,y2,..y q ]; Calculate the correlation between two vectors shifted by 12 periods in a cycle; perform sliding translation 12 times, each time moving forward one period, to obtain a new Y sequence. Calculate the correlation between the 12 sequences and the H sequence respectively. The translation period corresponding to the maximum correlation value is the lag period. If the lag period is greater than the preset threshold, it is determined that there is a risk anomaly. Step 4.3: If any of steps 4.1 and 4.2 identifies that the enterprise has risk anomalies, then the enterprise is determined to have risk anomalies.

8. A risk identification system for key nodes of the industrial chain, characterized in that: include: The industrial chain network construction module is used to obtain industrial chain node information based on predefined classification rules, establish an industrial chain network according to the upstream and downstream relationships of industrial chain nodes, and determine the industry and enterprise information of each industrial chain node based on the patent database and industrial and commercial information database; A correlation calculation module is used to determine the static correlation between the nodes in the industrial chain based on the upstream and downstream relationships of the nodes in the industrial chain and the input-output table of the entire industry, and to determine the dynamic correlation between the nodes in the industrial chain based on the power consumption data; A key node identification module, used to construct node characteristics, power consumption characteristics and resilience characteristics of the industrial chain nodes according to the static correlation and dynamic correlation between the industrial chain nodes, and determine the key nodes in the industrial chain network based on the node characteristics, power consumption characteristics and resilience characteristics; The risk identification module is used to perform risk monitoring on the identified key nodes of the industrial chain, wherein the risk monitoring includes historical electricity consumption monitoring and enterprise production lag monitoring, and identify risk nodes based on the risk monitoring results.

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the method for identifying risks of key nodes in the industrial chain according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for identifying risks of key nodes in the industrial chain described in any one of claims 1 to 7 are implemented.

Images