Enterprise risk management system based on big data analysis
Through big data analysis and BP neural network model, an enterprise risk management system is built, and the problem of insufficient risk identification and early warning in traditional methods is solved, and accurate assessment and timely warning of enterprise risks is achieved to ensure the stable development of the enterprise.
Patent Information
- Application Number
- CN202510542506.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-28
- Publication Date
- 2025-08-12
AI Technical Summary
Traditional corporate risk management methods rely on experience and intuition, lack scientificity and systemicity, and cannot identify and warn of potential risks in a timely manner, resulting in companies facing the risk of major losses.
Adopt an enterprise risk management system based on big data analysis, and define key risk indicators by collecting multi-source operation data, establishing a time series model for sales prediction, building a BP neural network model for risk prediction, and setting a risk threshold for early warning.
It has achieved accurate assessment of corporate risks, provided reliable decision-making basis, timely discovered the risk of sales decline, comprehensively considered a variety of risk factors, improved the scientificity and accuracy of decision-making, and ensured the stable development of the company.
Smart Images
Figure CN120471436A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of enterprise management, and in particular to an enterprise risk management system based on big data analysis. Background Art
[0002] With the rapid development of information technology, big data has become a crucial tool for corporate decision-making and management. In today's information explosion, businesses are faced with the need to accumulate and process massive amounts of data. This data includes not only internal production, sales, and financial data, but also external information such as supply chains, customer behavior, and market trends. Extracting valuable information from these massive datasets is crucial to enhancing a company's competitiveness.
[0003] Enterprises face various risks during their operations. If these risks are not identified and managed in a timely manner, they may cause significant losses to the enterprise or even endanger its survival. Traditional risk management methods often rely on experience and intuition, lack scientificity and systematicness, cannot provide early warning of potential risks, and lack effective risk management. Summary of the Invention
[0004] The present invention aims to solve the technical problems existing in the prior art and provides an enterprise risk management system based on big data analysis.
[0005] The present invention solves the above-mentioned technical problems with the following technical solutions: An enterprise risk management system based on big data analysis, comprising:
[0006] Risk Definition Module: This module collects the enterprise's multi-source operational data, correlates and integrates the data from different data sources, and defines the enterprise's key risk indicators based on the multi-source operational data;
[0007] Sales Forecasting Module: Collects sales data within the company's sales cycle and organizes it into a time series model. A sales forecast sequence is obtained through simple moving average forecasting. The trend of the time series model and the sales forecast sequence is compared to determine the company's sales risk.
[0008] Risk prediction module: obtains relevant data on enterprise operation risk prediction as input variables, establishes an enterprise operation risk prediction model, and conducts a comprehensive prediction of the enterprise's operation risks;
[0009] Early warning judgment module: Set risk thresholds, judge the enterprise's operational risks based on the probability of enterprise operational risk status output by the risk prediction model, and send early warnings in a timely manner.
[0010] In a preferred embodiment, the risk definition module obtains multi-source operational data of the enterprise, including production data, sales data, procurement data, and inventory data, pre-processes the collected multi-source operational data, and associates and integrates the data from different data sources. Based on the multi-source operational data, the module defines the key risk indicators of the enterprise, including production risk, sales risk, procurement risk, and inventory risk.
[0011] Production risk is obtained through equipment failure rate and production plan completion rate. The equipment failure rate is calculated by recording the cumulative number of equipment failures during the company's production cycle and counting the total operating time of the equipment during the company's production cycle. The specific calculation formula is as follows:
[0012]
[0013] Where F represents the cumulative number of equipment failures, T total It represents the total operating time that should have been used. The equipment failure rate reflects the frequency of equipment failures during the enterprise's production cycle. The production plan completion rate is calculated by obtaining the production plan output formulated by the enterprise and counting the actual completed output. The specific calculation formula is as follows:
[0014]
[0015] Among them, Q plan Indicates the planned production output, Q actual The actual output is calculated based on the production plan completion rate, which can show the completion of the company's production tasks within the production cycle. A production plan completion rate close to 100% indicates that the production plan is well executed and the production risk is low.
[0016] Sales risk is measured through customer churn rate and new customer acquisition rate. To measure customer churn rate, the company needs to obtain the total number of customers it has, and then determine the number of customers it has lost. Customers lost are those who have not made a repeat purchase. The customer churn rate is calculated based on the total number of customers and the number of lost customers. The specific calculation formula is as follows:
[0017]
[0018] Among them, C lost Represents the total number of customers, C begin The new customer acquisition rate represents the number of lost customers. It is necessary to count the number of newly developed customers who have completed their first transaction within the company, and estimate the total number of potential customers in the company's market through market research. The new customer acquisition rate is calculated based on the number of first-time transaction customers and the total number of potential customers. The specific calculation formula is as follows:
[0019]
[0020] Among them, C new represents the number of customers who made their first transaction, C potential Represents the total number of potential customers. The new customer acquisition rate reflects the company's ability to develop new markets and attract new customers. A high new customer acquisition rate indicates that the company has great development potential in the market and low sales risk.
[0021] Procurement risk is captured through the purchase price volatility and supplier delivery on-time rate. The purchase price volatility needs to record the price of each purchase of key raw materials. Assuming that a total of n purchases are made during the enterprise's production cycle, the average purchase price of the raw materials during the enterprise's production cycle is calculated using the following formula:
[0022]
[0023] in, represents the average purchase price of raw materials, P i Represents the price of each purchase of key raw materials. The purchase price volatility is calculated based on the average purchase price. The specific calculation formula is as follows:
[0024]
[0025] Volatility represents the fluctuation rate of the purchase price. The supplier acquisition rate is calculated by obtaining the total number of supplier deliveries and then determining the number of on-time deliveries. The specific calculation formula is as follows:
[0026]
[0027] Among them, punctuality represents the supplier's delivery punctuality rate, D on Indicates the number of on-time deliveries, D total The supplier's total number of deliveries. The supplier's on-time delivery rate reflects the importance the supplier places on delivery. A high on-time rate indicates a high supplier reliability and a low risk of production interruption due to untimely raw material supply.
[0028] Inventory risk is obtained through the inventory out-of-stock rate and inventory backlog rate. The inventory out-of-stock rate is calculated by counting the total number of inventory product categories that are out of stock and the total number of inventory product types of the company. The specific calculation formula is as follows:
[0029]
[0030] Among them, S out-stock Indicates the number of inventory product types that are out of stock, Stotal It represents the total number of product types in the company's inventory. The inventory quantity exceeding 1.5 times the average monthly sales volume is considered as overstock. The number of inventory products that meet the overstock criteria and the total inventory quantity are counted. The inventory overstock ratio is calculated based on the number of overstock products and the total inventory quantity. The specific calculation formula is as follows:
[0031]
[0032] Among them, I over Indicates the number of overstocked products, I total Representing the total number of inventory products, the inventory backlog rate indicator reflects the severity of the company's inventory backlog. A high backlog rate indicates that the company's inventory management efficiency is low, the company's occupancy costs are high, and the inventory products face the risk of obsolescence and depreciation, which increases the company's inventory risk and operating costs.
[0033] In a preferred embodiment, the sales forecast module collects the sales volume data of the enterprise during the sales cycle, and organizes the collected sales volume data into a time series model. The time series model of the sales volume data can be expressed as Y = {y1, y2, ..., y x}, where y i Denotes the sales volume corresponding to the i-th time point, x represents the total number of data points. Use data visualization tools to plot the sales volume time series model into a line graph. The basic trend of sales volume can be preliminarily judged through the line graph. The future sales volume can be predicted using the simple moving average forecasting method. The moving average period k is determined. Starting from the k-th data point, the moving average of each time point is calculated in sequence as the predicted value. For time point i, the specific calculation formula of the moving average is as follows:
[0034]
[0035] Among them, y j represents the actual sales data at the jth time point in the time series Y, It represents the moving average forecast value of the i-th time point in the time series, i represents the time point sequence in the time series, and the average value of the first k actual sales is taken as the forecast sales volume at the current time point. The moving average sequence is obtained based on the calculation of the moving average value. As a simple prediction result of future sales, compare the time series model Y of sales data and the sales series of future sales By observing the trend of the forecast series, we can preliminarily determine that there is a risk of sales decline.
[0036] In a preferred embodiment, the risk prediction module obtains relevant data of enterprise operation risk prediction from the risk definition module, including production risk X1, sales risk X2, procurement risk X3, inventory risk X4 and whether the enterprise is in a high-risk state P, high risk is recorded as 1 and low risk is recorded as 0, and X1, X2, and X3 are used as input variables x i , taking P as the output variable, based on the above content, a supply chain disruption risk prediction model is constructed. The specific steps are as follows:
[0037] S1. Normalize the relevant data of enterprise operation risk prediction;
[0038] S2. Determine the structure of the neural network, including the number of nodes in the input layer, hidden layer, and output layer, and initialize the weights and thresholds in the neural network;
[0039] S3. Perform forward propagation, perform weighted and offset processing on the original input data, and calculate the input value of each neuron in the hidden layer through the connection weights and input values between the input layer neurons and the hidden layer neurons, as well as the threshold of the hidden layer neurons. The specific calculation formula is as follows:
[0040]
[0041] Among them, net j represents the input of the hidden layer, w ij represents the connection weight between the i-th neuron in the input layer and the j-th neuron in the hidden layer, x i represents the input value of the i-th neuron in the input layer, θ j Represents the threshold of the jth neuron in the hidden layer. The activation function is used to perform nonlinear transformation on the input of the hidden layer. The specific calculation formula is as follows:
[0042] H j =f(net j )
[0043] Among them, f represents the activation function, H j Represents the output value of the jth neuron in the hidden layer, and the output value of the hidden layer neuron H i , the input value of each neuron in the output layer is calculated through the connection weight between the hidden layer and the output layer and the threshold of the output layer neurons. The specific calculation formula is as follows:
[0044]
[0045] Among them, net k represents the input value of the kth neuron in the output layer, N represents the number of neurons in the hidden layer, and w jkrepresents the connection weight between the jth neuron in the hidden layer and the kth neuron in the output layer, θ k Represents the threshold of the kth neuron in the output layer. The activation function is used again to process the input of the output layer to obtain the final predicted output value. k =f(net k );
[0046] S4. Calculation error: The deviation between the predicted output value and the actual output value is calculated through the mean square error. The deviation is used to measure the accuracy of the model prediction. It reflects the degree of deviation between the current model and the actual situation when predicting the enterprise operation risk. The specific calculation formula is as follows:
[0047]
[0048] Among them, E represents the error between the predicted output and the actual output, l represents the number of neurons in the output layer, and d k represents the actual output value of the kth neuron in the output layer, H k Represents the predicted output value of the kth neuron in the output layer. By minimizing this error, the model can continuously adjust its own parameters and improve its prediction ability;
[0049] S5, back propagation and parameter update: Based on the calculated error, the partial derivatives of the error with respect to the output layer weights and thresholds are used, and combined with the learning rate, the update amount of the output layer weights and thresholds is determined. The error information is used to adjust the model parameters. The specific calculation formula is as follows:
[0050]
[0051] Where Δw jk represents the output layer w jk The update amount, η represents the learning rate, Represents the error E versus weight w jk The partial derivative of Δθ k represents the output layer threshold θ k The update amount, is the error E with respect to the threshold θ k The update amount of the hidden layer weights and thresholds is determined by the partial derivative of the error with respect to the hidden layer weights and thresholds and the learning rate. The specific calculation formula is as follows:
[0052]
[0053] Where Δw ij represents the hidden layer weight w ij The update amount, Represents the error E versus weight w ij The partial derivative of Δθ j represents the hidden layer threshold θj The update amount, Represents the error E versus threshold θ j The partial derivative of
[0054] S6. Update weights and thresholds: Apply the calculated weight and threshold update amounts to the current weights and thresholds to update the model parameters. The specific calculation formula is as follows:
[0055] w ij (t+1)=w ij (t)+Δw ij
[0056] θ j (t+1)=θ j (t)+Δθ j
[0057] w jk (t+1)=w jk (t)+Δw jk
[0058] θ k (t+1)=θ k (t)+Δθ k
[0059] Among them, w ij (t+1),θ j (t+1), w jk (t+1),θ k (t+1) represents the weight between the input layer and the hidden layer, the hidden layer threshold, the weight between the hidden layer and the output layer, and the output layer threshold at the next iteration t+1;
[0060] S7. Repeat steps S3-S6 until the maximum number of iterations is reached.
[0061] In a preferred embodiment, the early warning judgment module sets a risk warning threshold based on the probability P of the enterprise operation being in a high-risk state output by the risk prediction module, combined with the enterprise's pre-set risk tolerance and business goals. When P≥T, it is determined that the current enterprise operation risk exceeds the acceptable range, and the system automatically triggers the early warning mechanism, which transmits it to relevant personnel by triggering a system pop-up window.
[0062] The beneficial effects of the present invention are: by collecting detailed data on production, sales, inventory, etc. and conducting in-depth analysis, the present invention can more accurately assess risk conditions, avoid subjective conjectures, provide a reliable basis for corporate decision-making, and improve the scientific nature and accuracy of decision-making. It uses BP neural network technology to construct a risk prediction model, and quickly output indicators such as the probability that the company's operation is in a high-risk state. It can detect risks in time when they first appear. By comparing the sales volume time series model and the forecast sequence, the risk of sales decline can be discovered as early as possible, so that the company has sufficient time to formulate a response strategy, integrate data from all aspects of the company, and comprehensively consider various risk factors such as production, sales, procurement, inventory, and their interrelationships, avoiding the limitations of single-factor evaluation, thereby conducting a full-scale and comprehensive assessment of the company's operating risks, helping companies to fully understand their own risk conditions, arrange preventive measures in advance, and ensure the stability and sustainable development of the company. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 is a flow chart of the present invention;
[0064] Figure 2 This is a system block diagram of the present invention. DETAILED DESCRIPTION
[0065] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.
[0066] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the described features. In the description of this application, "plurality" means two or more, unless otherwise specifically specified.
[0067] In the description of this application, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art will recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in this application.
[0068] like Figure 1-2 This embodiment provides: an enterprise risk management system based on big data analysis, including:
[0069] Risk Definition Module: This module collects the enterprise's multi-source operational data, correlates and integrates the data from different data sources, and defines the enterprise's key risk indicators based on the multi-source operational data;
[0070] In this embodiment, the risk definition module needs to be specifically explained. The risk definition module obtains multi-source operation data of the enterprise, including production data, sales data, procurement data, and inventory data, pre-processes the collected multi-source operation data, and associates and integrates the data from different data sources. Based on the multi-source operation data, the module defines the key risk indicators of the enterprise, including production risk, sales risk, procurement risk, and inventory risk.
[0071] Production risk is obtained through equipment failure rate and production plan completion rate. The equipment failure rate is calculated by recording the cumulative number of equipment failures during the company's production cycle and counting the total operating time of the equipment during the company's production cycle. The specific calculation formula is as follows:
[0072]
[0073] Where F represents the cumulative number of equipment failures, T total It represents the total operating time that should have been used. The equipment failure rate reflects the frequency of equipment failures during the enterprise's production cycle. The production plan completion rate is calculated by obtaining the production plan output formulated by the enterprise and counting the actual completed output. The specific calculation formula is as follows:
[0074]
[0075] Among them, Q plan Indicates the planned production output, Q actualThe actual output is calculated based on the production plan completion rate, which can show the completion of the company's production tasks within the production cycle. A production plan completion rate close to 100% indicates that the production plan is well executed and the production risk is low.
[0076] Sales risk is measured through customer churn rate and new customer acquisition rate. To measure customer churn rate, the company needs to obtain the total number of customers it has, and then determine the number of customers it has lost. Customers lost are those who have not made a repeat purchase. The customer churn rate is calculated based on the total number of customers and the number of lost customers. The specific calculation formula is as follows:
[0077]
[0078] Among them, C lost Represents the total number of customers, C begin The new customer acquisition rate represents the number of lost customers. It is necessary to count the number of newly developed customers who have completed their first transaction within the company, and estimate the total number of potential customers in the company's market through market research. The new customer acquisition rate is calculated based on the number of first-time transaction customers and the total number of potential customers. The specific calculation formula is as follows:
[0079]
[0080] Among them, C new represents the number of customers who made their first transaction, C potential Represents the total number of potential customers. The new customer acquisition rate reflects the company's ability to develop new markets and attract new customers. A high new customer acquisition rate indicates that the company has great development potential in the market and low sales risk.
[0081] Procurement risk is captured through the purchase price volatility and supplier delivery on-time rate. The purchase price volatility needs to record the price of each purchase of key raw materials. Assuming that a total of n purchases are made during the enterprise's production cycle, the average purchase price of the raw materials during the enterprise's production cycle is calculated using the following formula:
[0082]
[0083] in, represents the average purchase price of raw materials, P i Represents the price of each purchase of key raw materials. The purchase price volatility is calculated based on the average purchase price. The specific calculation formula is as follows:
[0084]
[0085] Volatility represents the fluctuation rate of the purchase price. The supplier acquisition rate is calculated by obtaining the total number of supplier deliveries and then determining the number of on-time deliveries. The specific calculation formula is as follows:
[0086]
[0087] Among them, punctuality represents the supplier's delivery punctuality rate, D on Indicates the number of on-time deliveries, D total The supplier's total number of deliveries. The supplier's on-time delivery rate reflects the importance the supplier places on delivery. A high on-time rate indicates a high supplier reliability and a low risk of production interruption due to untimely raw material supply.
[0088] Inventory risk is obtained through the inventory out-of-stock rate and inventory backlog rate. The inventory out-of-stock rate is calculated by counting the total number of inventory product categories that are out of stock and the total number of inventory product types of the company. The specific calculation formula is as follows:
[0089]
[0090] Among them, S out-stock Indicates the number of inventory product types that are out of stock, S total It represents the total number of product types in the company's inventory. The inventory quantity exceeding 1.5 times the average monthly sales volume is considered as overstock. The number of inventory products that meet the overstock criteria and the total inventory quantity are counted. The inventory overstock ratio is calculated based on the number of overstock products and the total inventory quantity. The specific calculation formula is as follows:
[0091]
[0092] Among them, I over Indicates the number of overstocked products, I total Representing the total number of inventory products, the inventory backlog rate indicator reflects the severity of the company's inventory backlog. A high backlog rate indicates that the company's inventory management efficiency is low, the company's occupancy costs are high, and the inventory products face the risk of obsolescence and depreciation, which increases the company's inventory risk and operating costs.
[0093] It should be noted that when associating sales data with production data and inventory data, the order number is the key association field. For the association of procurement data and inventory data, the material number and purchase batch number are used as the association basis.
[0094] Sales Forecasting Module: Collects sales data within the company's sales cycle and organizes it into a time series model. A sales forecast sequence is obtained through simple moving average forecasting. The trend of the time series model and the sales forecast sequence is compared to determine the company's sales risk.
[0095] In this embodiment, the sales forecast module needs to be specifically explained. The sales forecast module collects the sales volume data of the enterprise during the sales cycle and organizes the collected sales volume data into a time series model. The time series model of the sales volume data can be expressed as Y = {y1, y2, ..., y x}, where y i Denotes the sales volume corresponding to the i-th time point, x represents the total number of data points. Use data visualization tools to plot the sales volume time series model into a line graph. The basic trend of sales volume can be preliminarily judged through the line graph. The future sales volume can be predicted using the simple moving average forecasting method. The moving average period k is determined. Starting from the k-th data point, the moving average of each time point is calculated in sequence as the predicted value. For time point i, the specific calculation formula of the moving average is as follows:
[0096]
[0097] Among them, y j represents the actual sales data at the jth time point in the time series Y, It represents the moving average forecast value of the i-th time point in the time series, i represents the time point sequence in the time series, and the average value of the first k actual sales is taken as the forecast sales volume at the current time point. The moving average sequence is obtained based on the calculation of the moving average value. As a simple prediction result of future sales, compare the time series model Y of sales data and the sales series of future sales By observing the trend of the forecast series, we can preliminarily determine that there is a risk of sales decline.
[0098] It should be noted that sales data should cover multiple complete business cycles as much as possible, for example, data covering several years, to reflect fluctuations in different seasons and years. The horizontal axis of the sales time series model is time, and the vertical axis is sales. By observing the line graph, the general trend of sales over time can be intuitively analyzed, such as whether it is continuously rising or falling, or there are obvious seasonal fluctuations, cyclical changes, etc. The basic trend of sales can be preliminarily judged through the sales time series model, providing an intuitive basis for subsequent simple predictions and risk judgments. When using the simple moving average prediction method, if quarterly data is used, k=4 means calculating the average of the past four quarters to predict the next sales volume. Here, k is selected according to the data characteristics and actual business conditions.
[0099] Sales volume trend forecast is based on observing the trend of the forecast sequence. If the forecast sales volume shows a continuous downward trend, for example, the forecast value is decreasing in multiple consecutive forecast cycles, and the forecast value decreases by more than a certain percentage compared with the historical average sales volume, if the forecast value is set to be lower than Among them, r represents the risk threshold ratio set by the enterprise based on factors such as risk tolerance. For example, when r = 0.2, it means a decrease of 20%. It can be preliminarily judged that there is a risk of sales decline. Calculations based on sales forecasts indicate that there may be changes in consumer preferences, market saturation, product quality decline, and lack of product innovation. Sales risk directly affects the cash flow and profitability of the enterprise. Poor sales can easily lead to capital depletion, making it impossible for the enterprise to operate normally. When a company monitors a decline in sales, it should reform its current sales methods and technological innovation.
[0100] Risk prediction module: obtains relevant data on enterprise operation risk prediction as input variables, establishes an enterprise operation risk prediction model, and conducts a comprehensive prediction of the enterprise's operation risks;
[0101] In this embodiment, the risk prediction module needs to be specifically explained. The risk prediction module obtains relevant data of enterprise operation risk prediction from the risk definition module, including production risk X1, sales risk X2, procurement risk X3, inventory risk X4, and whether the enterprise is in a high-risk state P. High risk is recorded as 1 and low risk is recorded as 0. X1, X2, and X3 are used as input variables x i , taking P as the output variable, based on the above content, a supply chain disruption risk prediction model is constructed. The specific steps are as follows:
[0102] S1. Normalize the relevant data of enterprise operation risk prediction;
[0103] S2. Determine the structure of the neural network, including the number of nodes in the input layer, hidden layer, and output layer, and initialize the weights and thresholds in the neural network;
[0104] S3. Perform forward propagation, perform weighted and offset processing on the original input data, and calculate the input value of each neuron in the hidden layer through the connection weights and input values between the input layer neurons and the hidden layer neurons, as well as the threshold of the hidden layer neurons. The specific calculation formula is as follows:
[0105]
[0106] Among them, net j represents the input of the hidden layer, w ij represents the connection weight between the i-th neuron in the input layer and the j-th neuron in the hidden layer, x i represents the input value of the i-th neuron in the input layer, θj Represents the threshold of the jth neuron in the hidden layer. The activation function is used to perform nonlinear transformation on the input of the hidden layer. The specific calculation formula is as follows:
[0107] H j =f(net j )
[0108] Among them, f represents the activation function, H j Represents the output value of the jth neuron in the hidden layer, and the output value of the hidden layer neuron H i , the input value of each neuron in the output layer is calculated through the connection weight between the hidden layer and the output layer and the threshold of the output layer neurons. The specific calculation formula is as follows:
[0109]
[0110] Among them, net k represents the input value of the kth neuron in the output layer, N represents the number of neurons in the hidden layer, and w jk represents the connection weight between the jth neuron in the hidden layer and the kth neuron in the output layer, θ k Represents the threshold of the kth neuron in the output layer. The activation function is used again to process the input of the output layer to obtain the final predicted output value. k =f(net k );
[0111] S4. Calculation error: The deviation between the predicted output value and the actual output value is calculated through the mean square error. The deviation is used to measure the accuracy of the model prediction. It reflects the degree of deviation between the current model and the actual situation when predicting the enterprise operation risk. The specific calculation formula is as follows:
[0112]
[0113] Among them, E represents the error between the predicted output and the actual output, l represents the number of neurons in the output layer, and d k represents the actual output value of the kth neuron in the output layer, H k Represents the predicted output value of the kth neuron in the output layer. By minimizing this error, the model can continuously adjust its own parameters and improve its prediction ability;
[0114] S5, back propagation and parameter update: Based on the calculated error, the partial derivatives of the error with respect to the output layer weights and thresholds are used, and combined with the learning rate, the update amount of the output layer weights and thresholds is determined. The error information is used to adjust the model parameters. The specific calculation formula is as follows:
[0115]
[0116] Where Δwjk represents the output layer w jk The update amount, η represents the learning rate, Represents the error E versus weight w jk The partial derivative of Δθ k represents the output layer threshold θ k The update amount, is the error E with respect to the threshold θ k The update amount of the hidden layer weights and thresholds is determined by the partial derivative of the error with respect to the hidden layer weights and thresholds and the learning rate. The specific calculation formula is as follows:
[0117]
[0118]
[0119] Where Δw ij represents the hidden layer weight w ij The update amount, Represents the error E versus weight w ij The partial derivative of Δθ j represents the hidden layer threshold θ j The update amount, Represents the error E versus threshold θ j The partial derivative of
[0120] S6. Update weights and thresholds: Apply the calculated weight and threshold update amounts to the current weights and thresholds to update the model parameters. The specific calculation formula is as follows:
[0121] w ij (t+1)=w ij (t)+Δw ij
[0122] θ j (t+1)=θ j (t)+Δθ j
[0123] w jk (t+1)=w jk (t)+Δw jk
[0124] θ k (t+1)=θ k (t)+Δθ k
[0125] Among them, w ij (t+1),θ j (t+1), w jk (t+1),θ k(t+1) represents the weight between the input layer and the hidden layer, the hidden layer threshold, the weight between the hidden layer and the output layer, and the output layer threshold at the next iteration t+1;
[0126] S7. Repeat steps S3-S6 until the maximum number of iterations is reached.
[0127] It should be noted that in the forward propagation, assuming that the input data contains four variables: production risk, sales risk, procurement risk, and inventory risk, the forward propagation process can combine these variables in a nonlinear way to explore potential characteristic patterns that may affect the risk of enterprise operations. For example, when it is found that the production plan completion rate in production risk is lower than a certain threshold, the enterprise operation may face a higher risk. The nonlinear transformation can enable the neural network to learn more complex data relationships, which is helpful for modeling the complex problem of enterprise operation risk. Enterprise operation risk may be affected by the combined influence of multiple nonlinear related factors. The hidden layer output after activation function processing can better To capture these complex relationships, when the activation function is used again to process the input of the output layer in the calculation from the hidden layer to the output layer, the predicted output value will be used to compare with the actual output value to evaluate the accuracy of the model. In the enterprise operation risk prediction, the output value may represent relevant indicators such as the predicted probability that the enterprise is in a high-risk operation state. For example, an output value of 0.8 may mean that the model predicts that the probability of the enterprise operation being in a high-risk state is 80%. Based on this, the enterprise can formulate corresponding risk response strategies in advance, such as optimizing production processes, adjusting sales strategies, strengthening inventory management or replenishing cash flows, etc., to reduce the adverse effects of potential risks and ensure the stable operation and sustainable development of the enterprise.
[0128] Early warning judgment module: Set risk thresholds, determine the enterprise's operational risk based on the enterprise's operational risk status probability output by the risk prediction model, and send early warnings in a timely manner;
[0129] In this embodiment, what needs to be specifically explained is the early warning judgment module. The early warning judgment module sets a risk early warning threshold based on the probability P of the enterprise operation being in a high-risk state output by the risk prediction module, combined with the enterprise's pre-set risk tolerance and business goals. When P≥T, it is determined that the current enterprise operation risk exceeds the acceptable range, and the system automatically triggers the early warning mechanism, which transmits it to relevant personnel by triggering a system pop-up window.
[0130] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.
[0131] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0132] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0133] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.
[0134] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0135] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0136] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.
Claims
1. An enterprise risk management system based on big data analysis, characterized in that: include: Risk Definition Module: This module collects the enterprise's multi-source operational data, correlates and integrates the data from different data sources, and defines the enterprise's key risk indicators based on the multi-source operational data; Sales Forecasting Module: Collects sales data within the company's sales cycle and organizes it into a time series model. A sales forecast sequence is obtained through simple moving average forecasting. The trend of the time series model and the sales forecast sequence is compared to determine the company's sales risk. Risk prediction module: obtains relevant data on enterprise operation risk prediction as input variables, establishes an enterprise operation risk prediction model, and conducts a comprehensive prediction of the enterprise's operation risks; Early warning judgment module: Set risk thresholds, judge the enterprise's operational risks based on the probability of enterprise operational risk status output by the risk prediction model, and send early warnings in a timely manner.
2. The enterprise risk management system based on big data analysis according to claim 1, characterized in that: The risk definition module obtains the enterprise's multi-source operation data, including production data, sales data, procurement data and inventory data, pre-processes the collected multi-source operation data, associates and integrates data from different data sources, and defines the enterprise's key risk indicators based on the multi-source operation data, including production risk, sales risk, procurement risk and inventory risk.
3. The enterprise risk management system based on big data analysis according to claim 2, characterized in that: Production risk is obtained through equipment failure rate and production plan completion rate. The equipment failure rate is calculated by recording the cumulative number of equipment failures during the company's production cycle and counting the total operating time of the equipment during the company's production cycle. The specific calculation formula is as follows: Where F represents the cumulative number of equipment failures, T total It represents the total operating time that should have been used. The equipment failure rate reflects the frequency of equipment failures during the enterprise's production cycle. The production plan completion rate is calculated by obtaining the production plan output formulated by the enterprise and counting the actual completed output. The specific calculation formula is as follows: Among them, Q plan Indicates the planned production output, Q actual It indicates the actual output completed. The calculation of the production plan completion rate can show the completion status of the company's production tasks within the production cycle.
4. The enterprise risk management system based on big data analysis according to claim 2, characterized in that: Sales risk is measured through customer churn rate and new customer acquisition rate. To measure customer churn rate, the company needs to obtain the total number of customers it has, and then determine the number of customers it has lost. Customers lost are those who have not made a repeat purchase. The customer churn rate is calculated based on the total number of customers and the number of lost customers. The specific calculation formula is as follows: Among them, C lost Represents the total number of customers, C begin The new customer acquisition rate represents the number of lost customers. It is necessary to count the number of newly developed customers who have completed their first transaction within the company, and estimate the total number of potential customers in the company's market through market research. The new customer acquisition rate is calculated based on the number of first-time transaction customers and the total number of potential customers. The specific calculation formula is as follows: Among them, C new represents the number of customers who made their first transaction, C potential Represents the total number of potential customers. The new customer acquisition rate reflects the company's ability to develop new markets and attract new customers.
5. The enterprise risk management system based on big data analysis according to claim 2, characterized in that: Procurement risk is captured through the purchase price volatility and supplier delivery on-time rate. The purchase price volatility needs to record the price of each purchase of key raw materials. Assuming that a total of n purchases are made during the enterprise's production cycle, the average purchase price of the raw materials during the enterprise's production cycle is calculated using the following formula: in, represents the average purchase price of raw materials, P i Represents the price of each purchase of key raw materials. The purchase price volatility is calculated based on the average purchase price. The specific calculation formula is as follows: Volatility represents the fluctuation rate of the purchase price. The supplier acquisition rate is calculated by obtaining the total number of supplier deliveries and then determining the number of on-time deliveries. The specific calculation formula is as follows: Among them, punctuality represents the supplier's delivery punctuality rate, D on Indicates the number of on-time deliveries, D total Represents the total number of deliveries from the supplier.
6. The enterprise risk management system based on big data analysis according to claim 2, characterized in that: Inventory risk is obtained through the inventory out-of-stock rate and inventory backlog rate. The inventory out-of-stock rate is calculated by counting the total number of inventory product categories that are out of stock and the total number of inventory product types of the company. The specific calculation formula is as follows: Among them, S out-stock Indicates the number of inventory product types that are out of stock, S total It represents the total number of product types in the company's inventory. The inventory quantity exceeding 1.5 times the average monthly sales volume is considered as overstock. The number of inventory products that meet the overstock criteria and the total inventory quantity are counted. The inventory overstock ratio is calculated based on the number of overstock products and the total inventory quantity. The specific calculation formula is as follows: Among them, I over Indicates the number of overstocked products, I total Indicates the total number of inventory products. The inventory backlog ratio indicator reflects the severity of the company's inventory backlog, and a high backlog ratio.
7. The enterprise risk management system based on big data analysis according to claim 1, characterized in that: The sales forecast module collects the sales data of the enterprise during the sales cycle and organizes the collected sales data into a time series model. The time series model of the sales data can be expressed as Y = {y1, y2, ..., y x }, where y i Denotes the sales volume corresponding to the i-th time point, x represents the total number of data points. Use data visualization tools to plot the sales volume time series model into a line graph. The basic trend of sales volume can be preliminarily judged through the line graph. The future sales volume can be predicted using the simple moving average forecasting method. The moving average period k is determined. Starting from the k-th data point, the moving average of each time point is calculated in sequence as the predicted value. For time point i, the specific calculation formula of the moving average is as follows: Among them, y j represents the actual sales data at the jth time point in the time series Y, It represents the moving average forecast value of the i-th time point in the time series, i represents the time point sequence in the time series, and the average value of the first k actual sales is taken as the forecast sales volume at the current time point. The moving average sequence is obtained based on the calculation of the moving average value. As a simple prediction result of future sales, compare the time series model Y of sales data and the sales series of future sales By observing the trend of the forecast series, we can preliminarily determine that there is a risk of sales decline.
8. The enterprise risk management system based on big data analysis according to claim 1, characterized in that: The risk prediction module obtains relevant data of enterprise operation risk prediction from the risk definition module, including production risk X1, sales risk X2, procurement risk X3, inventory risk X4 and whether the enterprise is in a high-risk state P. High risk is recorded as 1 and low risk is recorded as 0. X1, X2, and X3 are used as input variables x i , taking P as the output variable, based on the above content, a supply chain disruption risk prediction model is constructed. The specific steps are as follows: S1. Normalize the relevant data of enterprise operation risk prediction; S2. Determine the structure of the neural network, including the number of nodes in the input layer, hidden layer, and output layer, and initialize the weights and thresholds in the neural network; S3. Perform forward propagation, perform weighted and offset processing on the original input data, and calculate the input value of each neuron in the hidden layer through the connection weights and input values between the input layer neurons and the hidden layer neurons, as well as the threshold of the hidden layer neurons. The specific calculation formula is as follows: Among them, net j represents the input of the hidden layer, w ij represents the connection weight between the i-th neuron in the input layer and the j-th neuron in the hidden layer, x i represents the input value of the i-th neuron in the input layer, θ j Represents the threshold of the jth neuron in the hidden layer. The activation function is used to perform nonlinear transformation on the input of the hidden layer. The specific calculation formula is as follows: H j =f(not j ) Among them, f represents the activation function, H j Represents the output value of the jth neuron in the hidden layer, and the output value of the hidden layer neuron H i , the input value of each neuron in the output layer is calculated through the connection weight between the hidden layer and the output layer and the threshold of the output layer neurons. The specific calculation formula is as follows: Among them, net k represents the input value of the kth neuron in the output layer, N represents the number of neurons in the hidden layer, and w jk represents the connection weight between the jth neuron in the hidden layer and the kth neuron in the output layer, θ k Represents the threshold of the kth neuron in the output layer. The activation function is used again to process the input of the output layer to obtain the final predicted output value. k =f(net k ); S4. Calculation error: The deviation between the predicted output value and the actual output value is calculated through the mean square error. The deviation is used to measure the accuracy of the model prediction. It reflects the degree of deviation between the current model and the actual situation when predicting the enterprise operation risk. The specific calculation formula is as follows: Among them, E represents the error between the predicted output and the actual output, l represents the number of neurons in the output layer, and d k represents the actual output value of the kth neuron in the output layer, H k Represents the predicted output value of the kth neuron in the output layer. By minimizing this error, the model can continuously adjust its own parameters and improve its prediction ability; S5, back propagation and parameter update: Based on the calculated error, the partial derivatives of the error with respect to the output layer weights and thresholds are used, and combined with the learning rate, the update amount of the output layer weights and thresholds is determined. The error information is used to adjust the model parameters. The specific calculation formula is as follows: Where Δw jk represents the output layer w jk The update amount, η represents the learning rate, Represents the error E versus weight w jk The partial derivative of Δθ k represents the output layer threshold θ k The update amount, is the error E with respect to the threshold θ k The update amount of the hidden layer weights and thresholds is determined by the partial derivative of the error with respect to the hidden layer weights and thresholds and the learning rate. The specific calculation formula is as follows: Where Δw ij represents the hidden layer weight w ij The update amount, Represents the error E versus weight w ij The partial derivative of Δθ j represents the hidden layer threshold θ j The update amount, Represents the error E versus threshold θ j The partial derivative of S6. Update weights and thresholds: Apply the calculated weight and threshold update amounts to the current weights and thresholds to update the model parameters. The specific calculation formula is as follows: w ij (t+1)=w ij (t)+Δw ij i j (t+1)=θ j (t)+Δθ j w jk (t+1)=w jk (t)+Δw jk i k (t+1)=θ k (t)+Δθ k Among them, w ij (t+1),θ j (t+1), w jk (t+1),θ k (t+1) represents the weight between the input layer and the hidden layer, the hidden layer threshold, the weight between the hidden layer and the output layer, and the output layer threshold at the next iteration t+1; S7. Repeat steps S3-S6 until the maximum number of iterations is reached.
9. The enterprise risk management system based on big data analysis according to claim 1, characterized in that: The early warning judgment module sets a risk warning threshold based on the probability P of the enterprise operation being in a high-risk state output by the risk prediction module, combined with the enterprise's pre-set risk tolerance and business goals. When P≥T, it is determined that the current enterprise operation risk exceeds the acceptable range, and the system automatically triggers the early warning mechanism, which transmits it to relevant personnel by triggering a system pop-up window.