Digital economic risk assessment method and device based on computer statistics

Abstract

The invention discloses a digital economic risk assessment method and device based on computer statistics, and the method comprises the steps: S1, collecting enterprise economic data and market environment real-time information data, and carrying out the data cleaning; s2, dividing enterprise cycles according to the enterprise economic data in S1; s3, extracting enterprise internal stage features and market features according to the enterprise cycle in S2; s4, combining the enterprise internal stage features and the market features to construct an enterprise periodic knowledge graph; s5, defining risk indexes according to the enterprise internal stage features and the market features; s6, constructing a risk assessment model in combination with the enterprise periodic knowledge graph and the risk indexes; s7, analyzing a risk propagation path and evaluating an influence degree through a knowledge graph; s8, carrying out risk early warning according to a risk assessment result, and proposing countermeasures and suggestions in combination with a risk propagation path and an influence range; risk assessment is carried out through enterprise economic data and market environment real-time information in combination with enterprise cycle division, feature extraction and knowledge graph construction.

Description

Technical Field

[0001] The present invention relates to the technical field of a data processing system or method specifically suitable for administrative, commercial, financial, management, supervisory or forecasting purposes, and in particular to the technical field of financial analysis, specifically a digital economic risk assessment method and device based on computer statistics. Background Art

[0002] Digital economy refers to a series of economic activities that use digital knowledge and information as key production factors, modern information networks as important carriers, and the effective use of information and communication technologies as an important driving force for improving efficiency and optimizing economic structure. In the era of digital economy, data has become the core asset of enterprises. Through efficient data processing and analysis, enterprises can optimize resource allocation, improve operational efficiency, and innovate business models. As important participants in the digital economy, enterprises' economic activities such as supply chain management, financing, and investment all involve a large amount of data, so the risks they face are more complex and diverse.

[0003] As an important participant in the digital economy, enterprises involve a large amount of data in their economic activities. In the digital economy era, data is the core element of enterprise operations and decision-making. With the rapid development of the digital economy, the amount of data involved in enterprise economic activities has increased dramatically, the data sources have become more extensive, and the data types have become more diverse, resulting in risk assessment of the digital economy becoming a necessary means for enterprise operations and decision-making.

[0004] Traditional risk assessment methods often rely on the experience and subjective judgment of experts, lack scientificity and objectivity, and are unable to cope with the complex and diverse data types and rapidly changing market environment in the digital economy era. In addition, traditional methods often focus on a single financial indicator or market indicator, ignoring the complex interaction between the internal operating conditions of the enterprise and the market environment, resulting in risk assessment results that are not comprehensive and accurate. In addition, traditional risk assessment uses a large amount of data for comparative evaluation, lacks intuitive feeling, and is not conducive to corporate management and decision makers to quickly grasp the risk situation and make corresponding decisions.

[0005] Therefore, it is necessary to improve the digital economy risk assessment method and device based on computer statistics in the prior art to solve the above problems. Summary of the invention

[0006] The present invention overcomes the shortcomings of the prior art and provides a digital economy risk assessment method and device based on computer statistics, aiming to solve the problems that traditional risk assessment methods in the prior art lack scientificity and objectivity, are difficult to handle the complex and diverse data types and rapidly changing market environment in the digital economy era, the assessment results are not comprehensive and accurate, and it is not convenient for enterprise management and decision makers to quickly grasp the risk status and make corresponding decisions.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is: a digital economic risk assessment method and device based on computer statistics, comprising:

[0008] S1. Collect enterprise economic data and real-time information data of market environment, and perform data cleaning;

[0009] S2, divide the enterprise cycle according to the economic data of enterprises in S1;

[0010] S3, extracting the internal stage characteristics and market characteristics of the enterprise based on the enterprise cycle in S2;

[0011] S4. Construct a knowledge graph of enterprise cycles based on the internal stage characteristics of the enterprise and the market characteristics;

[0012] S5. Define risk indicators based on the internal stage characteristics of the enterprise and market characteristics;

[0013] S6. Construct a risk assessment model by combining the enterprise cycle knowledge graph and risk indicators;

[0014] S7. Analyze the risk propagation path and evaluate the impact through knowledge graph.

[0015] In a preferred embodiment of the present invention, in step S1, the enterprise economic data includes: transaction volume, transaction frequency, enterprise growth, market share, revenue growth rate, and cost structure; the real-time information data of the market environment includes: policy changes, industry trends, competitor dynamics, macroeconomic indicators, and market indexes.

[0016] In a preferred embodiment of the present invention, in step S2, the business cycle includes: start-up phase, growth phase, maturity phase, and decline phase.

[0017] In a preferred embodiment of the present invention, in step S2, K-means clustering is performed on the standardized enterprise economic data; Among them, k = 4 is the number of clusters, namely the start-up stage, growth stage, maturity stage, and decline stage; C i is the i-th type of enterprise economic data, x∈C i is the data point of the i-th category of enterprise economic data after standardization, θ i It is the center point of the i-th category of enterprise economic data;

[0018] For each cluster, the ARIMA time series forecasting model is used to predict the possible future development trend of the enterprise.

[0019] in, AR reflects the linear relationship between the target value at time t and the previous t-1 to tp target values; tis the current value of the enterprise economic data, b is a constant, p is the number of autoregressive terms of the enterprise economic data, r i is the autocorrelation coefficient of enterprise economic data, ε t is the error; MA reflects that there is a linear relationship between the target value at time t and the previous t-1~p error values. q is the sliding average of the enterprise economic data, θ i is the moving average coefficient of the enterprise economic data; combining the two, we get the time series prediction model:

[0020] In a preferred embodiment of the present invention, in step S3, the Pearson correlation coefficient is used to analyze the correlation between the internal characteristics of the enterprise and the market characteristics, and the correlation relationship between the characteristics is constructed. The calculation formula of the Pearson correlation coefficient r is: Among them, X i and Y i are the observed values ​​of enterprise internal characteristics and market characteristics, and are the average values ​​of the internal characteristics of the enterprise and the market characteristics, n is the number of observations, and i refers to the i-th one.

[0021] In a preferred embodiment of the present invention, in step S4, the step of constructing an enterprise cycle knowledge graph includes:

[0022] S41. Define knowledge graph nodes;

[0023] Define the business cycle nodes, including start-up, growth, maturity and decline;

[0024] S42, define knowledge graph edges;

[0025] S421, define the relationship edge between the enterprise cycle and the internal characteristics and market characteristics of the enterprise, determine the correlation strength between the internal characteristics and market characteristics of the enterprise and each cycle, and define the edge and the weight of the edge accordingly;

[0026] Calculate the Pearson correlation coefficient between each firm's internal characteristics and market characteristics and each firm's cycle and Among them, X and Y are the internal characteristics of the enterprise and the market characteristics respectively, Q1, Q2, Q3, Q4 are the start-up period, growth period, maturity period and decline period of the enterprise cycle; the weight of the corresponding relationship edge is w = |r|, which indicates the strength of the association between the characteristics and the cycle;

[0027] S422. Based on business logic and data relationships, define the relationship edges between internal enterprise characteristics and market characteristics; use the Pearson correlation coefficient in S3 to calculate the weights of the corresponding relationship edges;

[0028] S43. Construct a knowledge graph of the enterprise cycle based on the defined nodes, edges and edge weights.

[0029] In a preferred embodiment of the present invention, in step S5, the step of defining risk indicators includes:

[0030] S51. Determine the risk type of the enterprise cycle;

[0031] S52. For each risk type, corresponding risk indicators are constructed based on the internal characteristics of the enterprise and the market characteristics to quantify the risk;

[0032] S53. Standardize the quantified risk indicators.

[0033] In a preferred embodiment of the present invention, in step S6, the step of constructing a risk assessment model includes:

[0034] S61. Extract all risk-related features from the enterprise cycle knowledge graph and combine them with the risk indicators to form a comprehensive feature vector;

[0035] The knowledge graph feature vector Z contains n features, Z = [z1, z2, …, z n ] T , the characteristics include: internal characteristics of the enterprise, market characteristics and enterprise cycle; combined with the standardized 15 risk indicators R′, a comprehensive characteristic vector is obtained: S = [Z; R′];

[0036] S62. Use logistic regression method to build a risk assessment model;

[0037] Define the weight corresponding to the comprehensive feature vector β = [β0,β1,…β n+15 ];

[0038] The probability of risk occurring

[0039] In a preferred embodiment of the present invention, in step S7, for each edge, the possibility of propagation or the degree of influence thereof is calculated, and the overall weight of the path is calculated based on the weight of the edge; Among them, W path represents the total weight of the path, w i represents the weight of the i-th edge in the path;

[0040] The risk impact is calculated based on the path weight and the importance of the affected nodes: I node It indicates the degree to which a node is affected by the risk, and Path(node) indicates the set of all propagation paths passing through the node.

[0041] The present invention provides a digital economic risk assessment device based on computer statistics, comprising:

[0042] Data collection module, used to collect enterprise economic data and real-time information data of market environment;

[0043] Data cleaning module, used to clean the data from the data collection module and perform data standardization to ensure data accuracy and consistency;

[0044] The enterprise cycle division module is used to divide the enterprise cycle through the K-means clustering algorithm.

[0045] The time series forecasting module is used to predict the future development trend of the enterprise using the ARIMA time series forecasting model to assist in the accuracy of the enterprise cycle division results;

[0046] The feature extraction module is used to extract enterprise internal features and market features from enterprise economic data and market environment real-time information data according to the enterprise cycle division results;

[0047] The knowledge graph construction module is used to define the nodes, edges and edge weights of the knowledge graph and construct the enterprise cycle knowledge graph based on the extracted internal enterprise characteristics, market characteristics and enterprise cycle division results;

[0048] Risk indicator definition module, used to define risk indicators according to enterprise cycle and risk type;

[0049] Risk indicator quantification module, used to quantify corresponding risk indicators;

[0050] The risk assessment model building module is used to build a risk assessment model by combining the features and risk indicators extracted from the enterprise cycle knowledge graph;

[0051] The risk propagation path module is used to analyze the potential propagation paths of risks within the enterprise and between markets through knowledge graphs, and to assess the degree and scope of risk impact;

[0052] The risk warning module is used to set the risk warning threshold according to the risk assessment results and trigger the warning mechanism when the risk probability exceeds the threshold.

[0053] The present invention solves the defects existing in the background technology and has the following beneficial effects:

[0054] (1) The present invention collects and cleans enterprise economic data and real-time information on the market environment, combines enterprise cycle division, feature extraction and knowledge graph construction, defines targeted risk indicators, and constructs a comprehensive risk assessment model. By analyzing the risk propagation path and impact degree through the knowledge graph, accurate risk warning and response measures are achieved. It not only improves the scientificity and accuracy of digital economic risk assessment, but also provides enterprises with a full range of risk management perspectives from the inside out. Compared with the existing technology, it significantly enhances the risk resistance of enterprises in a complex economic environment.

[0055] (2) The present invention adopts K-means clustering algorithm and ARIMA time series prediction model based on enterprise economic data to divide enterprises into start-up period, growth period, maturity period and decline period, so as to understand the operating characteristics of enterprises at different stages. K-means clustering algorithm classifies the economic data of enterprises to determine the cycle of enterprises. ARIMA time series prediction model is used to predict the future development trend of enterprises and assist in verifying the accuracy of clustering results. The division of enterprise cycle is combined with the construction of risk assessment model, which provides time dimension analysis for risk assessment, so that risk assessment can be customized for the specific cycle of enterprises, thereby improving the pertinence and effectiveness of assessment.

[0056] (3) The present invention constructs an enterprise cycle knowledge graph by extracting internal enterprise and market characteristics and analyzing the correlation between these characteristics using the Pearson correlation coefficient. By extracting key features and analyzing their relationships, the present invention reveals the complex interactions between the internal and external environments of an enterprise, thereby better understanding the root causes of risks and enhancing the understanding of the operating conditions of an enterprise. In addition, the present invention provides an intuitive perspective for the analysis of risk propagation paths in the form of a knowledge graph.

[0057] (4) The present invention defines risk indicators through the risk types of the enterprise cycle. By quantifying the risk indicators, the abstract risk concept can be converted into actionable data, thereby achieving objective risk assessment. All risk-related features are extracted from the enterprise cycle knowledge graph and combined with the risk indicators to form a comprehensive feature vector. The risk assessment model is then constructed using the logistic regression method, thereby achieving quantitative risk assessment and improving the accuracy of risk prediction.

[0058] (5) The present invention analyzes the risk propagation path and evaluates the degree of impact through knowledge graphs. By calculating the overall weight of the path and the degree of risk impact of the nodes, it can identify the key paths and vulnerable links of risk diffusion, and demonstrate the risk propagation mechanism within the enterprise and in the market environment. This provides a new perspective for risk management and helps enterprises take more targeted risk prevention and control measures.

[0059] (6) The present invention sets warning thresholds based on risk assessment results, so that enterprises can promptly detect and respond to potential risks and avoid or mitigate negative impacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 is a flow chart of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0064] A digital economy risk assessment method based on computer statistics, comprising the steps of:

[0065] S1. Collect enterprise economic data and real-time information data of market environment, and perform data cleaning;

[0066] S2, divide the enterprise cycle according to the economic data of enterprises in S1;

[0067] S3, extracting the internal stage characteristics and market characteristics of the enterprise based on the enterprise cycle in S2;

[0068] S4. Construct a knowledge graph of enterprise cycles based on the internal stage characteristics of the enterprise and the market characteristics;

[0069] S5. Define risk indicators based on the internal stage characteristics of the enterprise and market characteristics;

[0070] S6. Construct a risk assessment model by combining the enterprise cycle knowledge graph and risk indicators;

[0071] S7. Analyze the risk propagation path and assess the impact through knowledge graph;

[0072] S8. Issue risk warnings based on risk assessment results, and propose response measures based on the risk transmission path and impact scope.

[0073] Enterprise economic data refers to various data that reflect the economic activities and financial status of an enterprise. It directly reflects the operating conditions and financial health of the enterprise and is an important basis for assessing the digital economic risks of the enterprise. Real-time information data on the market environment refers to external environmental factors that affect the operation of the enterprise. Enterprise economic data and market environment data comprehensively reflect the operating conditions and external environment of the enterprise from both internal and external dimensions, making enterprise economic data and real-time information data on the market environment the basis for conducting digital economic risk assessment.

[0074] Specifically, in step S1, the enterprise economic data includes: transaction volume, transaction frequency, enterprise growth, market share, revenue growth rate, and cost structure;

[0075] Transaction volume and frequency reflect the business activity and market participation of an enterprise; enterprise growth, including growth in revenue, profit, assets, etc., reflects the development speed and potential of an enterprise; market share reflects the position and competitiveness of an enterprise in the market; revenue growth rate shows the growth trend of an enterprise's revenue; and cost structure reflects the cost structure and profitability of an enterprise.

[0076] The reason for selecting these data is that they can comprehensively and objectively reflect the operating conditions, financial status and market position of enterprises, providing a solid foundation for digital economy risk assessment.

[0077] Specifically, in step S1, the real-time information data of the market environment includes: policy changes, industry trends, competitor dynamics, macroeconomic indicators, and market indexes;

[0078] Policy changes reflect the direction and speed of industry development; industry trends reflect the direction and speed of industry development; competitor dynamics help companies formulate targeted competitive strategies; macroeconomic indicators, including GDP and CPI, reflect the overall economic environment; market indices, including stock market indices and industry indices, reflect the overall market trend and volatility;

[0079] The reason for selecting these real-time information data on the market environment is that they can provide companies with important information about the market environment, industry trends and competitors, helping companies better understand the impact of external environmental factors on their operations and digital economic risks.

[0080] In step S1, data cleaning is performed on the enterprise economic data and the real-time information data of the market environment. Data cleaning refers to the process of checking data consistency, processing invalid values ​​and missing values, etc., including: identifying inaccurate or incomplete data, and then repairing, deleting or replacing these data; through cleaning, errors and anomalies in the data can be eliminated, the accuracy and reliability of the data can be improved, and the accuracy and stability of the model can be improved.

[0081] In step S1, the data cleaning step includes:

[0082] S11. Classify the enterprise economic data and market environment real-time information data into numerical data and categorical data according to data type, check whether there are missing values, and fill in the missing values;

[0083] Numerical data include: transaction volume, transaction frequency, enterprise growth, market share, revenue growth rate, macroeconomic indicators, market index; categorical data include: cost structure, policy changes, industry trends, competitor dynamics;

[0084] For numerical data, linear interpolation is used to supplement missing data: Among them, d lost Indicates missing values, t lost The time when the missing value is indicated, d lost+1 and d lost-1 Indicates the missing value followed by a non-missing data value and the previous non-missing data value, t lost+1 and t lost-1 represents the time corresponding to the non-missing data value;

[0085] For categorical data, the majority filling method is used to supplement the missing data and fill in the data with the highest number of occurrences.

[0086] S12, performing outlier detection and correction on the filled enterprise economic data and market environment real-time information data;

[0087] For numerical data, the IQR method is used to identify outliers: IQR = Q3-Q1, where Q3 is the third quartile and Q1 is the first quartile. Outlier determination: outlier < Q1-1.5×IQR or outlier > Q3+1.5×IQR; for detected outliers, delete or replace them with the mean or median;

[0088] S13. Perform data standardization processing on the revised enterprise economic data and market environment real-time information data;

[0089] Assign status to classified data according to different situations and business logic:

[0090] The following assignments are a specific embodiment:

[0091] The cost structure is based on the ratio of fixed costs to variable costs, and assigns values ​​to different ratio ranges:

[0092] Very high fixed cost ratio = 5, fixed costs account for more than 70% of total costs;

[0093] High fixed cost ratio = 4, the ratio of fixed cost to total cost is between 60% and 70%;

[0094] Fixed cost and variable cost balance = 3, the ratio of fixed cost to variable cost is roughly equal, that is, each accounts for about 50%;

[0095] High variable cost ratio = 2, the ratio of variable cost to total cost is between 30% and 40%;

[0096] Very high variable cost ratio = 1, the ratio of variable cost to total cost is less than 30%;

[0097] Policy changes are based on the degree of impact of the policy on the enterprise and are assigned values ​​according to the size of the impact;

[0098] Very favorable = 5, the policy has a significant positive impact on business operations, resulting in revenue growth of more than 20%;

[0099] Favorable = 4, the policy has a positive impact on business operations, resulting in revenue growth between 10%-20%;

[0100] Neutral = 3, the policy has no significant impact on business operations, and revenue growth is between 0%-10%;

[0101] Unfavorable = 2, the policy has a negative impact on business operations, resulting in a revenue reduction of between 10% and 20%;

[0102] Very unfavorable = 1, the policy has a significant negative impact on the company's operations, resulting in a reduction in revenue of more than 20%;

[0103] Industry trends, which are assigned values ​​according to growth rate or market potential;

[0104] Rapid growth = 5, the industry's annual growth rate exceeds 20%

[0105] Stable growth = 4, the industry's annual growth rate is between 10%-20%;

[0106] Stable = 3, the industry's annual growth rate is between 0%-10%;

[0107] Slow growth = 2, the industry's annual growth rate is between -10% and 0%;

[0108] Recession = 1, the industry's annual growth rate is less than -10%;

[0109] Competitor dynamics, which uses competitors’ market performance as a benchmark and assigns values ​​based on changes in market share or competitiveness;

[0110] Significant increase in market share = 5, the competitor's market share increased by more than 10% in the most recent fiscal year;

[0111] Market share increase = 4, the competitor's market share increased between 5% and 10% in the most recent fiscal year;

[0112] Market share stability = 3, the competitor's market share did not change by more than ±5% in the most recent fiscal year;

[0113] Market share reduction = 2, the competitor's market share has decreased by between 5% and 10% in the most recent fiscal year;

[0114] Significant Market Share Loss = 1; The competitor’s market share has decreased by more than 10% in the most recent fiscal year.

[0115] After the assignment, the categorical data is transformed into numerical data and subjected to Z-score standardization; Among them, x is the standardized value, a is the original data, μ is the mean value, and σ is the standard deviation;

[0116] The business cycle refers to the different stages that a company goes through in the course of its development. These stages reflect the company's growth, maturity and possible decline. The division of business cycles helps to understand the company's operating characteristics, challenges and opportunities at different stages, thereby providing support for corporate decision-making.

[0117] Specifically, in step S2, the enterprise cycle includes: start-up phase, growth phase, maturity phase, and decline phase;

[0118] The start-up phase is when a company is just established, usually with limited funds and products or services not yet widely accepted by the market; the growth phase is when the company gradually gains market recognition after the start-up phase, and sales revenue and profits begin to grow rapidly; the mature phase is when the company has achieved a certain market share and scale of operation, and the growth rate slows down but remains stable; the decline phase is when the company's sales revenue and profits begin to decline due to intensified market competition, technological upgrades, or changes in consumer demand.

[0119] Perform K-means clustering on the standardized enterprise economic data; Among them, k = 4 is the number of clusters, namely the start-up stage, growth stage, maturity stage, and decline stage; C i is the i-th type of enterprise economic data, x∈C i is the data point of the i-th category of enterprise economic data after standardization, It is the center point of the i-th category of the enterprise's economic data; the K-means clustering algorithm classifies the enterprise's economic data to determine the cycle in which the enterprise is located, and the ARIMA time series prediction model is used to predict the future development trend of the enterprise and assist in verifying the accuracy of the clustering results;

[0120] For each cluster, the ARIMA time series forecasting model is used to predict the possible future development trend of the enterprise to assist in the interpretation and verification of the clustering results;

[0121] in, AR reflects the linear relationship between the target value at time t and the previous t-1 to tp target values; t is the current value of the enterprise economic data, b is a constant, p is the number of autoregressive terms of the enterprise economic data, r i is the autocorrelation coefficient of enterprise economic data, ε t is the error; MA reflects that there is a linear relationship between the target value at time t and the previous t-1~p error values. q is the sliding average of the enterprise economic data, θ i is the moving average coefficient of the enterprise economic data; combining the two, we get the time series prediction model:

[0122] The basis for judging the start-up stage: high revenue growth rate and low market share; for example: annual revenue growth rate exceeds 30%, and revenue growth exceeds 10% for two consecutive quarters, and market share is 20% lower than the industry average, and the absolute value is less than 5%;

[0123] The basis for judging the growth stage: high and stable revenue growth rate and rising market share trend; specifically, for example: annual revenue growth rate is between 20%-30%, and the revenue growth rate has been stable at 5%-10% for four consecutive quarters; market share has increased by at least 2% each year for two consecutive fiscal years, and the current market share is 20% higher than the industry average.

[0124] The criteria for determining the maturity stage are: the revenue growth rate and market share growth rate slow down, the cost structure is optimized, and the profit margin is stable. Specifically, for example, the annual revenue growth rate is less than 20%, and the revenue growth rate does not exceed 5% for four consecutive quarters; the ratio of fixed costs to variable costs tends to be balanced; the profit margin remains above the industry average, and the fluctuation does not exceed ±5%;

[0125] Recession stage: Revenue growth rate and market share show a downward trend, cost structure deteriorates, and profit margins decline; specifically, for example: the annual revenue growth rate is negative for two consecutive years, and the revenue has decreased by more than 5% month-on-month for four consecutive quarters, the cost-effectiveness ratio deteriorates, the proportion of fixed costs is too high compared to variable costs or variable costs are poorly controlled, resulting in a cost / revenue ratio higher than the industry average, and the profit margin has declined for two consecutive fiscal years and is lower than the industry average.

[0126] There is a time sequence in the corporate cycle. Through the above numerical judgment basis, combined with K-means clustering and ARIMA model, the numerical value of the next stage can be predicted and the cycle of the next stage can be judged. This can divide the current corporate cycle of the company and provide a basis for subsequent risk assessment.

[0127] In step S3, the extraction of internal enterprise characteristics and market characteristics is based on the enterprise cycle and related enterprise economic data and real-time information data of the market environment. The internal enterprise characteristics mainly reflect the operation, finance, management and other aspects of the enterprise, while the market characteristics mainly describe the performance of the enterprise in the market and the status of the market environment.

[0128] From corporate economic data, we can extract characteristics that reflect the company's financial and operating conditions, such as revenue growth rate, cost structure, and profit margin. Combined with the company cycle, we can analyze changes in the company's internal management and operating strategies at different stages. For example, the start-up phase focuses more on innovation and R&D, while the mature phase focuses more on cost control and market maintenance. From real-time information data on the market environment, we can extract characteristics that reflect market conditions, such as market share, competitor dynamics, and industry trends. According to changes in the company cycle and market environment, we can analyze the company's market performance and strategy adjustments at different stages. For example, the company may actively expand its market share during the growth phase, but may face a reduction in market share during the recession phase.

[0129] Specifically, in step S3,

[0130] The internal characteristics of start-up companies include: high R&D investment, low profit margin, unoptimized cost structure, and low operational efficiency; the market characteristics of start-up companies include: low market share, low brand awareness, and sensitive response to competitors' dynamics;

[0131] Internal characteristics of growth-stage enterprises include: high revenue growth rate, improved profit margin, improved operational efficiency, and stable cost structure; Market characteristics of growth-stage enterprises include: rapid growth in market share, gradually increasing brand influence, and close attention to competitor dynamics;

[0132] The internal characteristics of mature enterprises include: stable revenue growth rate, optimized cost structure, high operational efficiency, and stable profit margin; the characteristics of mature markets include: stable market share and high brand awareness.

[0133] The internal characteristics of enterprises in the recession stage include: declining revenue growth rate, declining profit margins, deteriorating cost structure, and reduced operational efficiency; the characteristics of the mature market include: reduced market share, weakened brand influence, and slow response to competitors' dynamics.

[0134] The Pearson correlation coefficient is used to analyze the correlation between internal enterprise characteristics and market characteristics, to build the correlation between characteristics, and to lay the foundation for the construction of the knowledge graph. The calculation formula of the Pearson correlation coefficient r is: Among them, X i and Y i are the observed values ​​of enterprise internal characteristics and market characteristics, and are the average values ​​of the internal characteristics of the enterprise and the market characteristics, respectively; n is the number of observations, and i refers to the i-th observation; the correlation coefficient ranges from -1 to 1. A value close to 1 indicates a strong positive correlation, a value close to -1 indicates a strong negative correlation, and a value close to 0 indicates no correlation or nonlinear correlation.

[0135] The knowledge graph organizes information through nodes and edges, showing the correlation between internal enterprise and market characteristics in different cycles. The knowledge graph provides an intuitive and structured information representation for risk assessment, helps to deeply understand the operating conditions of the enterprise, and provides rich information for the construction of risk assessment models.

[0136] Specifically, in step S4, the steps of constructing the enterprise cycle knowledge graph include:

[0137] S41. Define knowledge graph nodes;

[0138] Define the business cycle nodes, including start-up, growth, maturity and decline;

[0139] According to the internal characteristics and market characteristics of the enterprise extracted by S3, corresponding internal characteristic nodes and market characteristic nodes are defined for each enterprise cycle;

[0140] S42, define knowledge graph edges;

[0141] S421, define the relationship edge between the enterprise cycle and the internal characteristics and market characteristics of the enterprise, determine the correlation strength between the internal characteristics and market characteristics of the enterprise and each cycle, and define the edge and the weight of the edge accordingly;

[0142] Calculate the Pearson correlation coefficient between each firm's internal characteristics and market characteristics and each firm's cycle and Among them, X and Y are the internal characteristics of the enterprise and the market characteristics respectively, Q1, Q2, Q3, Q4 are the start-up period, growth period, maturity period and decline period of the enterprise cycle; the weight of the corresponding relationship edge is w = |r|, which indicates the strength of the association between the characteristics and the cycle.

[0143] Edge weights are used to represent the strength of association between internal enterprise characteristics or market characteristics and enterprise cycle nodes. Since the Pearson correlation coefficient can measure the linear correlation between two variables, it can be used as the basis for calculating edge weights. The absolute value of the Pearson correlation coefficient is taken because the knowledge graph only cares about the strength of the association between the two, not whether it is positively or negatively correlated.

[0144] S422. Based on business logic and data relationships, define the relationship edges between internal enterprise characteristics and market characteristics; use the Pearson correlation coefficient in S3 to calculate the weights of the corresponding relationship edges;

[0145] S43. Construct a knowledge graph of the enterprise cycle according to the defined nodes, edges and edge weights;

[0146] Risk indicators are the core input of risk assessment models and directly determine the accuracy of model predictions;

[0147] Specifically, in step S5, the risk indicator definition step includes:

[0148] S51. Determine the risk type of the enterprise cycle;

[0149] Start-up risks include: capital liquidity risk, product innovation risk;

[0150] Growth stage risks include: market expansion risk, supply chain risk;

[0151] Risks in the maturity stage include: risks of intensified market competition and brand aging;

[0152] Recession risks include: financial crisis risk, transition risk;

[0153] S52. For each risk type, corresponding risk indicators are constructed based on the internal characteristics of the enterprise and the market characteristics to quantify the risk;

[0154] Quantitative indicators of fund liquidity risk: Quantitative indicators of product innovation risk:

[0155] Market expansion risk quantitative indicators: Market expansion deviation = |Company revenue growth rate - Industry revenue growth rate|;

[0156]

[0157] Supply chain risk quantitative indicators:

[0158] Quantitative indicators of the risk of intensified market competition:

[0159] Quantitative indicators of brand aging risk:

[0160] Quantitative indicators of financial crisis risk:

[0161] Transformation risk quantitative indicators:

[0162] S53: Standardize the quantified risk indicators; the standardization method is consistent with the standardization method in step S1 Where R is the risk index, μ R and σ R are the mean and standard deviation of the risk indicator, respectively;

[0163] The risk assessment model outputs risk probability, providing a scientific basis for risk warning and the formulation of response measures;

[0164] Specifically, in step S6, the step of building a risk assessment model includes:

[0165] S61. Extract all risk-related features from the enterprise cycle knowledge graph and combine them with the risk indicators to form a comprehensive feature vector;

[0166] The knowledge graph feature vector Z contains n features, Z = [z1, z2, …, z n ] T , the characteristics include: internal characteristics of the enterprise, market characteristics and enterprise cycle; combined with the standardized 15 risk indicators R′, a comprehensive characteristic vector is obtained: S = [Z; R′];

[0167] S62. Use logistic regression method to build a risk assessment model;

[0168] Define the weight corresponding to the comprehensive feature vector β = [β0,β1,…β n+15 ];

[0169] The probability of risk occurring

[0170] When pi is close to 1, it indicates high risk, and when pi is close to 0, it indicates low risk;

[0171] Risk propagation path and impact analysis provides management with a detailed view of risk propagation, helping them understand the overall impact of risks and thus develop more comprehensive and effective response measures.

[0172] Specifically, in step S7, for each edge, the possibility of its propagation or the degree of its influence is calculated, and the overall weight of the path is calculated based on the weight of the edge (i.e., the strength of the correlation between the features); Among them, W path represents the total weight of the path, w i represents the weight of the i-th edge in the path;

[0173] The risk impact is calculated based on the path weight and the importance of the affected nodes: I node It indicates the degree to which a node is affected by the risk, and Path(node) indicates the set of all propagation paths passing through the node.

[0174] Specifically, in step S8, according to the evaluation result, a risk warning threshold T=δ×pi′ is set, where δ represents the risk tolerance. When the risk probability pi′ is greater than or equal to the threshold T, a warning is triggered.

[0175] A digital economic risk assessment device based on computer statistics, comprising:

[0176] Data collection module, used to collect enterprise economic data and real-time information data of market environment;

[0177] Data cleaning module, used to clean the data from the data collection module and perform data standardization to ensure data accuracy and consistency;

[0178] The enterprise cycle division module is used to divide the enterprise cycle through the K-means clustering algorithm.

[0179] The time series forecasting module is used to predict the future development trend of the enterprise using the ARIMA time series forecasting model to assist in the accuracy of the enterprise cycle division results;

[0180] The feature extraction module is used to extract enterprise internal features and market features from enterprise economic data and market environment real-time information data according to the enterprise cycle division results;

[0181] The knowledge graph construction module is used to define the nodes, edges and edge weights of the knowledge graph and construct the enterprise cycle knowledge graph based on the extracted internal enterprise characteristics, market characteristics and enterprise cycle division results;

[0182] Risk indicator definition module, used to define risk indicators according to enterprise cycle and risk type;

[0183] Risk indicator quantification module, used to quantify corresponding risk indicators;

[0184] The risk assessment model building module is used to build a risk assessment model by combining the features and risk indicators extracted from the enterprise cycle knowledge graph;

[0185] The risk propagation path module is used to analyze the potential propagation paths of risks within the enterprise and between markets through knowledge graphs, and to assess the degree and scope of risk impact;

[0186] The risk warning module is used to set the risk warning threshold according to the risk assessment results and trigger the warning mechanism when the risk probability exceeds the threshold;

[0187] The response suggestion module is used to put forward specific response measures.

[0188] The above is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.

Claims

1. A digital economy risk assessment method based on computer statistics, characterized in that: Includes steps: S1. Collect enterprise economic data and real-time information data of market environment, and perform data cleaning; S2, divide the enterprise cycle according to the economic data of enterprises in S1; S3, extracting the internal stage characteristics and market characteristics of the enterprise based on the enterprise cycle in S2; S4. Construct a knowledge graph of enterprise cycles based on the internal stage characteristics of the enterprise and the market characteristics; S5. Define risk indicators based on the internal stage characteristics of the enterprise and market characteristics; S6. Construct a risk assessment model by combining the enterprise cycle knowledge graph and risk indicators; S7. Analyze the risk propagation path and evaluate the impact through knowledge graph.

2. According to claim 1, a digital economy risk assessment method based on computer statistics is characterized by: In step S1, the enterprise economic data includes: transaction volume, transaction frequency, enterprise growth, market share, revenue growth rate, and cost structure; the real-time information data of the market environment includes: policy changes, industry trends, competitor dynamics, macroeconomic indicators, and market indexes.

3. According to claim 1, a digital economy risk assessment method based on computer statistics is characterized by: In step S2, the enterprise cycle includes: start-up phase, growth phase, maturity phase, and decline phase.

4. According to claim 1, a method for digital economic risk assessment based on computer statistics is characterized in that: In step S2, K-means clustering is performed on the standardized enterprise economic data; Among them, k = 4 is the number of clusters, namely the start-up stage, growth stage, maturity stage, and decline stage; C i is the i-th type of enterprise economic data, x∈C i is the data point of the i-th category of enterprise economic data after standardization, It is the center point of the i-th category of enterprise economic data; For each cluster, the ARIMA time series forecasting model is used to predict the possible future development trend of the enterprise. in, AR reflects the linear relationship between the target value at time t and the previous t-1 to tp target values; t is the current value of the enterprise economic data, b is a constant, p is the number of autoregressive terms of the enterprise economic data, r i is the autocorrelation coefficient of enterprise economic data, ε t is the error; MA reflects that there is a linear relationship between the target value at time t and the previous t-1~p error values. q is the sliding average of the enterprise economic data, θ i is the moving average coefficient of the enterprise economic data; combining the two, we get the time series prediction model:

5. According to claim 1, a method for digital economic risk assessment based on computer statistics is characterized in that: In step S3, the Pearson correlation coefficient is used to analyze the correlation between the internal characteristics of the enterprise and the market characteristics, and the correlation relationship between the characteristics is constructed. The calculation formula of the Pearson correlation coefficient r is: Among them, X i and Y i are the observed values ​​of enterprise internal characteristics and market characteristics, and are the average values ​​of the internal characteristics of the enterprise and the market characteristics, n is the number of observations, and i refers to the i-th one.

6. According to a computer statistics-based digital economy risk assessment method according to claim 1, it is characterized by: In step S4, the steps of constructing the enterprise cycle knowledge graph include: S41. Define knowledge graph nodes; Define the business cycle nodes, including start-up, growth, maturity and decline; S42, define knowledge graph edges; S421, define the relationship edge between the enterprise cycle and the internal characteristics and market characteristics of the enterprise, determine the correlation strength between the internal characteristics and market characteristics of the enterprise and each cycle, and define the edge and the weight of the edge accordingly; Calculate the Pearson correlation coefficient between each firm's internal characteristics and market characteristics and each firm's cycle and Among them, X and Y are the internal characteristics of the enterprise and the market characteristics respectively, Q1, Q2, Q3, Q4 are the start-up period, growth period, maturity period and decline period of the enterprise cycle; the weight of the corresponding relationship edge is w = |r|, which indicates the strength of the association between the characteristics and the cycle; S422. Based on business logic and data relationships, define the relationship edges between internal enterprise characteristics and market characteristics; use the Pearson correlation coefficient in S3 to calculate the weights of the corresponding relationship edges; S43. Construct a knowledge graph of the enterprise cycle based on the defined nodes, edges and edge weights.

7. The method for digital economic risk assessment based on computer statistics according to claim 1 is characterized in that: In step S5, the risk indicator definition step includes: S51. Determine the risk type of the enterprise cycle; S52. For each risk type, corresponding risk indicators are constructed based on the internal characteristics of the enterprise and the market characteristics to quantify the risk; S53. Standardize the quantified risk indicators.

8. The method for digital economic risk assessment based on computer statistics according to claim 1 is characterized by: In step S6, the step of constructing a risk assessment model includes: S61. Extract all risk-related features from the enterprise cycle knowledge graph and combine them with the risk indicators to form a comprehensive feature vector; The knowledge graph feature vector Z contains n features, Z = [z1, z2, …, z n ] T , the characteristics include: internal characteristics of the enterprise, market characteristics and enterprise cycle; combined with the standardized 15 risk indicators R′, a comprehensive characteristic vector is obtained: S = [Z; R′]; S62. Use logistic regression method to build a risk assessment model; Define the weight corresponding to the comprehensive feature vector β = [β0, β1, …β n+15 ]; The probability of risk occurring 9. The method for digital economic risk assessment based on computer statistics according to claim 1 is characterized by: In step S7, for each edge, the possibility of its propagation or the degree of its influence is calculated, and the overall weight of the path is calculated based on the weight of the edge; Among them, W path represents the total weight of the path, w i represents the weight of the i-th edge in the path; The risk impact is calculated based on the path weight and the importance of the affected nodes: I node It indicates the degree to which a node is affected by the risk, and Path(node) indicates the set of all propagation paths passing through the node.

10. A digital economy risk assessment device based on computer statistics, a device based on a digital economy risk assessment method based on computer statistics according to any one of claims 1 to 9, characterized in that: Data collection module, used to collect enterprise economic data and real-time information data of market environment; Data cleaning module, used to clean the data from the data collection module and perform data standardization to ensure data accuracy and consistency; The enterprise cycle division module is used to divide the enterprise cycle through the K-means clustering algorithm. The time series forecasting module is used to predict the future development trend of the enterprise using the ARIMA time series forecasting model to assist in the accuracy of the enterprise cycle division results; The feature extraction module is used to extract enterprise internal features and market features from enterprise economic data and market environment real-time information data according to the results of enterprise cycle division; The knowledge graph construction module is used to define the nodes, edges and edge weights of the knowledge graph and construct the enterprise cycle knowledge graph based on the extracted internal enterprise characteristics, market characteristics and enterprise cycle division results; Risk indicator definition module, used to define risk indicators according to enterprise cycle and risk type; Risk indicator quantification module, used to quantify corresponding risk indicators; The risk assessment model building module is used to build a risk assessment model by combining the features and risk indicators extracted from the enterprise cycle knowledge graph; The risk propagation path module is used to analyze the potential propagation paths of risks within the enterprise and between markets through knowledge graphs, and to assess the degree and scope of risk impact; The risk warning module is used to set the risk warning threshold according to the risk assessment results and trigger the warning mechanism when the risk probability exceeds the threshold.

Images