Consumer industry business closure risk early warning method, device and storage medium
By using multi-source data processing and fuzzy comprehensive evaluation, the system addresses the lag problem in traditional store closure risk assessment, enabling objective and accurate risk assessment and dynamic early warning for stores. It integrates market and consumer data into a risk assessment system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- RED SHIELD BIG DATA (BEIJING) CO LTD
- Filing Date
- 2026-05-07
- Publication Date
- 2026-07-24
AI Technical Summary
Traditional methods for assessing store closure risks rely on managerial experience or single financial indicators, lacking objective standards and failing to promptly capture external market threats, resulting in delayed risk warnings.
By employing multi-source data collection, cleaning, and standardization, combined with least squares method to fit demand curves, ARMA time series model, and analytic hierarchy process, a fuzzy comprehensive evaluation system is constructed to dynamically assess store risk levels and send early warnings.
It enables objective and accurate assessment of store risks, integrates macro market supply and demand, micro competitor and consumer sentiment data, and achieves dynamic and real-time risk monitoring and early warning.
Smart Images
Figure CN122453155A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of commercial data analysis technology, and in particular to a method, device and storage medium for early warning of store closure risks for enterprises in the consumer industry. Background Technology
[0002] In the consumer industry, businesses face the challenge of a rapidly changing market environment for survival and development. Intense market competition, rapidly shifting consumer preferences, and economic cycle fluctuations all exacerbate the operational risks of physical stores or brand outlets, leading to frequent store closures. Traditional methods for assessing store closure risk largely rely on managers' personal experience or single financial indicators (such as monthly sales or profit margins). These methods have the following shortcomings:
[0003] ① Relying on managers' experience to make judgments is easily influenced by personal perceptions and emotions, and lacks objective standards. Financial data usually reflects results that have already occurred, and is a lagging indicator. By the time risks are clearly reflected in financial statements, the best time for intervention has often been missed.
[0004] ② Focusing solely on one's own sales data fails to capture potential threats from external markets.
[0005] Therefore, the consumer industry urgently needs a method for early warning of store closure risks to help businesses take effective measures before risks materialize and ensure operational safety. Summary of the Invention
[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method, device, and storage medium for early warning of store closure risks for businesses in the consumer industry.
[0007] To achieve the above objectives, the present invention adopts the following technical solution:
[0008] The risk warning method for store closures in the consumer industry includes the following steps:
[0009] S1: Multi-source data acquisition;
[0010] S2: Clean the collected data, fill in missing values, remove duplicates, and standardize it;
[0011] S3: Use the least squares method to fit the demand curve, calculate the price elasticity of demand, determine the sensitivity of commodity demand to price, and assess market stability.
[0012] S4: Using the ARMA time series model, historical sales data is processed and modeled to predict future short-term and medium-term sales trends;
[0013] S5: Using the analytic hierarchy process, compare the importance of each risk indicator pairwise, construct a judgment matrix, and calculate and verify the weight of each indicator.
[0014] S6: Fuzzy comprehensive evaluation, with a preset set of comments, substitutes the actual value of each risk indicator into its membership function to obtain the membership vector of a single indicator, and then combines the weight vectors of all indicators with the membership matrix to obtain a comprehensive risk membership vector. Based on the principle of maximum membership, the final risk level of the store is determined.
[0015] S7: Dynamic early warning and tracking. When the system determines that the store's risk level has reached the set level, the early warning module automatically sends early warning information to the designated path.
[0016] S8: The system continuously tracks the latest operating data of the store and re-executes the analysis and evaluation process.
[0017] A risk warning device for store closures in the consumer industry, used to implement the above methods, including:
[0018] The data acquisition module is responsible for collecting data from multiple channels;
[0019] The data processing and analysis module cleans, organizes, and standardizes the collected data, removing noisy and duplicate data; and uses data analysis algorithms to extract information from the data.
[0020] The risk assessment module constructs a comprehensive risk assessment index system based on the results of data processing and analysis; assigns corresponding weights to each index, and uses fuzzy comprehensive evaluation method to integrate the evaluation results of each index to obtain a comprehensive score of store closure risk; a preset comment set is used to calculate and weight the sum of the membership functions of each risk index value to finally obtain a comprehensive risk membership vector, thereby determining the risk level of the store.
[0021] The early warning module is used to send early warning information, continuously track changes in the store's subsequent operating data, reassess the risk situation based on the new data, and update the early warning information in a timely manner.
[0022] The storage module is used to store the raw data collected, the data after processing and analysis, and historical records of risk assessment and early warning.
[0023] Preferably, the data acquisition module acquires data in the following ways: obtaining the total supply and demand data of clothing products in the business district in the past quarter from a professional data provider through an API interface; using web crawler technology to crawl user posts on the platform that mention the store and its competitors, and performing text sentiment analysis to quantify consumer enthusiasm indicators.
[0024] Preferably, the data processing and analysis module uses the elasticity coefficient formula for analyzing market supply and demand:
[0025]
[0026] Indicates the price elasticity of demand. It is the percentage change in demand. It represents the percentage change in price; by calculating the price elasticity of demand for various commodities over different time periods, we can determine the sensitivity of market demand to price changes and assess market stability.
[0027] Preferably, the data processing and analysis module combines a time series analysis model to fit and predict store sales data, as follows:
[0028]
[0029] in, For a moment Sales volume For constant terms, and For model parameters, This is a random error term used to predict future short- and medium-term sales trends.
[0030] Preferably, the data processing and analysis module includes a data cleaning and standardization unit, a market demand elasticity analysis unit, and a sales trend forecasting unit;
[0031] The data cleaning and standardization unit fills in missing values in numerical sales data by calculating the average of all non-null values in the field; for categorical variables, it fills in the mode; for duplicate data, it determines whether a data item is duplicated by setting a unique identifier or comparing the similarity of the data content. If a data item is duplicated, it retains only one complete record. After cleaning, it performs standardization transformation on all data.
[0032] Preferably, the market demand elasticity analysis unit collects data on different price levels of a certain commodity in the market and their corresponding sales volume within a set time period; and, under the condition that other factors remain unchanged, fits the demand curve equation using the least squares method. Where Q represents quantity demanded, P represents price, and a and b are parameters to be determined; using the least squares method, the sum of squared residuals between the actual observed values and the regression estimates is minimized to solve for the values of parameters a and b; subsequently, the price elasticity of demand coefficient is calculated. .
[0033] Preferably, the sales trend prediction unit uses ARMA in time series analysis to model and predict sales data; firstly, it performs a stationarity test on historical sales data; if the data is not stationary, it is made stationary through differencing; let the stationary sequence be... The order of the ARMA model is determined based on the characteristics of the autocorrelation function and the partial autocorrelation function. .
[0034] Preferably, the risk assessment module includes:
[0035] The indicator weight determination unit uses the analytic hierarchy process (AHP) to determine the weights of each risk indicator. First, a hierarchical model is constructed, with store closure risk as the target layer, market risk, operational risk, and consumer enthusiasm indicators as the criteria layer, and the sub-indicators under each criteria layer as the solution layer. Then, the relative importance of each element within the same layer to the elements of the previous layer is compared and scored pairwise to form a judgment matrix. The largest eigenvalue of the judgment matrix and its corresponding eigenvector are calculated, and after normalization, the weight vector of each element is obtained.
[0036] The fuzzy comprehensive evaluation unit first determines the set of comments and defines a corresponding membership function for each comment; then, it substitutes the actual observed values of each risk indicator into the corresponding membership function to calculate their respective membership vectors; finally, it performs a weighted synthesis operation based on the pre-determined indicator weight vector and membership matrix to obtain the comprehensive risk membership vector; and selects the comment with the highest membership degree as the final risk level determination result.
[0037] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.
[0038] The beneficial effects of this invention are as follows:
[0039] 1. This invention uses demand elasticity analysis and ARMA time series forecasting for data analysis, transforming qualitative market perceptions into quantitative indicators; in the risk assessment stage, the analytic hierarchy process (AHP) is used to determine weights and the fuzzy comprehensive evaluation method, which effectively reduces the bias of subjective judgment and makes the risk level determination more objective and accurate.
[0040] 2. This invention integrates macro market supply and demand, micro competitor dynamics, and social media sentiment data reflecting consumer emotions to construct a three-dimensional assessment dimension, making risk identification more comprehensive.
[0041] 3. This invention automatically reassesses the risk status based on the latest data and updates the early warning information, realizing dynamic and real-time monitoring of store risks, which makes it easier for managers to discover trends early and make quick decisions. Attached Figure Description
[0042] Figure 1 This is a flowchart of the consumer industry enterprise store closure risk early warning method proposed in this invention. Detailed Implementation
[0043] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.
[0044] Example 1:
[0045] The risk warning method for store closures in the consumer industry includes the following steps:
[0046] S1: Multi-source data acquisition;
[0047] S2: Clean the collected data, fill in missing values, remove duplicates, and standardize it;
[0048] S3: Use the least squares method to fit the demand curve, calculate the price elasticity of demand (Ed), determine the sensitivity of commodity demand to price, and assess market stability;
[0049] S4: Using the ARMA time series model, historical sales data is processed and modeled to predict future short-term and medium-term sales trends;
[0050] S5: Using the Analytic Hierarchy Process (AHP), compare the importance of each risk indicator (such as market risk, operational risk, etc.) pairwise, construct a judgment matrix, and calculate and verify the weight of each indicator.
[0051] S6: Fuzzy comprehensive evaluation, with a preset set of comments, substitutes the actual value of each risk indicator into its membership function to obtain the membership vector of a single indicator, and then combines the weight vectors of all indicators with the membership matrix to obtain a comprehensive risk membership vector. Based on the principle of maximum membership, the final risk level of the store is determined.
[0052] S7: Dynamic early warning and tracking. When the system determines that the store's risk level has reached the set level, the early warning module automatically sends early warning information to the designated path.
[0053] S8: The system continuously tracks the latest operating data of the store and re-executes the analysis and evaluation process.
[0054] Example 2:
[0055] Consumer goods businesses facing store closure risk warning devices include:
[0056] The data acquisition module is responsible for collecting data from multiple channels, including industry supply and demand data provided by market research institutions, consumer enthusiasm data reflected on social media platforms (such as the popularity of related topics, user sentiment, etc.), the store's own sales data, and the operating data of surrounding competitors.
[0057] The data processing and analysis module cleans, organizes, and standardizes the collected data, removing noise and duplicate data; and uses data analysis algorithms to extract valuable information from the data.
[0058] The risk assessment module constructs a comprehensive risk assessment index system based on the results of data processing and analysis; assigns corresponding weights to each index, and uses fuzzy comprehensive evaluation method to integrate the evaluation results of each index to obtain a comprehensive score of store closure risk; a preset comment set is used to calculate and weight the sum of the membership functions of each risk index value to finally obtain a comprehensive risk membership vector, thereby determining the risk level of the store.
[0059] The early warning module is used to send early warning information, continuously track changes in the store's subsequent operating data, reassess the risk situation based on the new data, and update the early warning information in a timely manner.
[0060] The storage module is used to store the raw data collected, the data after processing and analysis, and historical records of risk assessment and early warning.
[0061] The data collection module collects data in the following ways: it obtains the total supply and demand data of clothing products in the business district in the past quarter from a professional data provider through an API interface; it uses web crawler technology to crawl user posts on platforms such as Weibo and Xiaohongshu that mention the store and its competitors, and performs text sentiment analysis to quantify consumer enthusiasm indicators.
[0062] The data processing and analysis module uses an elasticity coefficient formula to analyze market supply and demand:
[0063]
[0064] Indicates the price elasticity of demand. It is the percentage change in demand. It represents the percentage change in price; by calculating the price elasticity of demand for various commodities over different time periods, we can determine the sensitivity of market demand to price changes and assess market stability.
[0065] The data processing and analysis module, in particular, uses a time series analysis model to fit and predict store sales data, as follows:
[0066]
[0067] in, For a moment Sales volume For constant terms, and For model parameters, This is a random error term used to predict future short- and medium-term sales trends.
[0068] The risk assessment module constructs a risk assessment indicator system, and the weights assigned to the indicators are determined using the analytic hierarchy process or the entropy weight method. The risk assessment module's preset evaluation set includes low risk, medium risk, and high risk. When the risk assessment module determines that a store is at a medium or high risk level, the early warning module sends an early warning message to a designated path.
[0069] The data processing and analysis module includes a data cleaning and standardization unit, a market demand elasticity analysis unit, and a sales trend forecasting unit.
[0070] The data cleaning and standardization unit fills in missing values in numerical sales data by calculating the average of all non-null values in the field; for categorical variables, it fills in the mode; for duplicate data, it determines whether the data is duplicate by setting a unique identifier or comparing the similarity of the data content. If the data is duplicate, it retains only one complete record. After cleaning, it performs standardization transformation on all data.
[0071] The market demand elasticity analysis unit collects data on different price levels of a certain commodity in the market and their corresponding sales volumes within a set time period; and, under other unchanged conditions, fits the demand curve equation using the least squares method. Where Q represents quantity demanded, P represents price, and a and b are parameters to be determined; using the least squares method, the sum of squared residuals between the actual observed values and the regression estimates is minimized to solve for the values of parameters a and b; subsequently, the price elasticity of demand coefficient is calculated. .
[0072] when When the price of a commodity is highly elastic, it indicates that price changes have a relatively large impact on the quantity demanded; when... When the price of a commodity is inelastic, it indicates that price changes have a relatively small impact on the quantity demanded; when the price of a commodity is inelastic, it indicates that the quantity demanded is relatively small. At that time, it is in a state of unit elasticity; by calculating and analyzing the elasticity coefficients of different time periods and different product categories, we can gain a deeper understanding of the characteristics and stability of market demand, and provide a basis for formulating reasonable pricing strategies and procurement plans.
[0073] The sales trend prediction unit employs ARMA (Archived Average Scale) in time series analysis for sales data modeling and prediction. First, it performs a stationarity test on historical sales data. If the data is not stationary, it is stabilized through differencing. Let the stabilized sequence be... The order of the ARMA model is determined based on the characteristics of the autocorrelation function and the partial autocorrelation function. ;
[0074] If the ACF tails and the PACF is truncated at lag p, then the AR(p) model is selected; if the ACF is truncated at lag q and the PACF tails, then the MA(q) model is selected; if both tail, then the ARMA(p,q) model is selected. The unknown parameters in the model are solved using the maximum likelihood estimation method. After obtaining the well-fitted model, it is applied to the short-term and medium-term forecasts of the future.
[0075] For example, given the monthly sales data for the past 12 months, and having established an ARMA(2,1) model through the steps described above, inputting the time index values for the next few months will yield the corresponding sales forecast values, thus helping stores to manage inventory and plan marketing in advance.
[0076] The risk assessment module includes:
[0077] The indicator weight determination unit uses the analytic hierarchy process (AHP) to determine the weight of each risk indicator. First, a hierarchical structure model is constructed, with store closure risk as the target layer, market risk, operational risk, and consumer enthusiasm indicators as the criteria layer, and the sub-indicators under each criteria layer as the solution layer. Then, the relative importance of each element in the same layer to the elements in the previous layer is compared and scored pairwise to form a judgment matrix.
[0078] For example, for the two criteria-level elements, market risk and operational risk, if their importance ratio to store closure risk is considered to be 4:6, then the values 4 and 6 (and other similar comparison results) are filled into the corresponding positions in the judgment matrix; the largest eigenvalue of the judgment matrix and its corresponding eigenvector are calculated, and after normalization, the weight vector of each element is obtained; a consistency test is performed, and the consistency index CI and consistency ratio CR are calculated. When CR < 0.1, the judgment matrix has satisfactory consistency; otherwise, the judgment matrix needs to be readjusted until it passes the test.
[0079] The fuzzy comprehensive evaluation unit first determines the set of comments and defines a corresponding membership function for each comment; then, it substitutes the actual observed values of each risk indicator into the corresponding membership function to calculate their respective membership vectors; finally, it performs a weighted synthesis operation based on the pre-determined indicator weight vector and membership matrix to obtain the comprehensive risk membership vector; according to the principle of maximum membership, the comment with the largest membership is selected as the final risk level determination result.
[0080] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A method for early warning of store closure risks in the consumer industry, characterized by: Includes the following steps: S1: Multi-source data acquisition; S2: Clean the collected data, fill in missing values, remove duplicates, and standardize it; S3: Use the least squares method to fit the demand curve, calculate the price elasticity of demand, determine the sensitivity of commodity demand to price, and assess market stability. S4: Using the ARMA time series model, historical sales data is processed and modeled to predict future short-term and medium-term sales trends; S5: Using the analytic hierarchy process, compare the importance of each risk indicator pairwise, construct a judgment matrix, and calculate and verify the weight of each indicator. S6: Fuzzy comprehensive evaluation, with a preset set of comments, substitutes the actual value of each risk indicator into its membership function to obtain the membership vector of a single indicator, and then combines the weight vectors of all indicators with the membership matrix to obtain a comprehensive risk membership vector. Based on the principle of maximum membership, the final risk level of the store is determined. S7: Dynamic early warning and tracking. When the system determines that the store's risk level has reached the set level, the early warning module automatically sends early warning information to the designated path. S8: The system continuously tracks the latest operating data of the store and re-executes the analysis and evaluation process.
2. A risk warning device for store closures in the consumer industry, characterized in that, To implement the method of claim 1, comprising: The data acquisition module is responsible for collecting data from multiple channels; The data processing and analysis module cleans, organizes, and standardizes the collected data, removing noisy and duplicate data; and uses data analysis algorithms to extract information from the data. The risk assessment module constructs a comprehensive risk assessment index system based on the results of data processing and analysis; assigns corresponding weights to each index, and uses fuzzy comprehensive evaluation method to integrate the evaluation results of each index to obtain a comprehensive score of store closure risk; a preset comment set is used to calculate and weight the sum of the membership functions of each risk index value to finally obtain a comprehensive risk membership vector, thereby determining the risk level of the store. The early warning module is used to send early warning information, continuously track changes in the store's subsequent operating data, reassess the risk situation based on the new data, and update the early warning information in a timely manner. The storage module is used to store the raw data collected, the data after processing and analysis, and historical records of risk assessment and early warning.
3. The consumer industry enterprise closure risk early warning device according to claim 2, characterized in that, The data acquisition module collects data in the following ways: it obtains the total supply and demand data of clothing products in the business district in the past quarter from professional data providers through API interfaces; it uses web crawler technology to crawl user posts on the platform that mention the store and its competitors, and performs text sentiment analysis to quantify consumer enthusiasm indicators.
4. The consumer industry enterprise closure risk early warning device according to claim 2, characterized in that, The data processing and analysis module uses an elasticity coefficient formula to analyze market supply and demand: , Indicates the price elasticity of demand. It is the percentage change in demand. It represents the percentage change in price; by calculating the price elasticity of demand for various commodities over different time periods, we can determine the sensitivity of market demand to price changes and assess market stability.
5. The consumer industry enterprise closure risk early warning device according to claim 4, characterized in that, The data processing and analysis module, combined with a time series analysis model, fits and predicts store sales data, as follows: in, For a moment Sales volume For constant terms, and For model parameters, This is a random error term used to predict future short- and medium-term sales trends.
6. The consumer industry enterprise closure risk early warning device according to claim 2, characterized in that, The data processing and analysis module includes a data cleaning and standardization unit, a market demand elasticity analysis unit, and a sales trend forecasting unit. The data cleaning and standardization unit fills in missing values in numerical sales data by calculating the average of all non-null values in the field; for categorical variables, it fills in the mode; for duplicate data, it determines whether a data item is duplicated by setting a unique identifier or comparing the similarity of the data content. If a data item is duplicated, it retains only one complete record. After cleaning, it performs standardization transformation on all data.
7. The consumer industry enterprise closure risk early warning device according to claim 6, characterized in that, The market demand elasticity analysis unit collects data on different price levels of a certain commodity in the market and their corresponding sales volumes within a set time period; and, under other unchanged conditions, fits the demand curve equation using the least squares method. Where Q represents quantity demanded, P represents price, and a and b are parameters to be determined; using the least squares method, the sum of squared residuals between the actual observed values and the regression estimates is minimized to solve for the values of parameters a and b; subsequently, the price elasticity of demand coefficient is calculated. .
8. The consumer industry enterprise store closure risk early warning device according to claim 6, characterized in that, The sales trend prediction unit uses ARMA in time series analysis to model and predict sales data. First, it performs a stationarity test on historical sales data; if the data is not stationary, it is stabilized through differencing. Let the stabilized sequence be... The order of the ARMA model is determined based on the characteristics of the autocorrelation function and the partial autocorrelation function. .
9. The consumer industry enterprise closure risk early warning device according to claim 2, characterized in that, The risk assessment module includes: The indicator weight determination unit uses the analytic hierarchy process (AHP) to determine the weights of each risk indicator. First, a hierarchical model is constructed, with store closure risk as the target layer, market risk, operational risk, and consumer enthusiasm indicators as the criteria layer, and the sub-indicators under each criteria layer as the solution layer. Then, the relative importance of each element within the same layer to the elements of the previous layer is compared and scored pairwise to form a judgment matrix. The largest eigenvalue of the judgment matrix and its corresponding eigenvector are calculated, and after normalization, the weight vector of each element is obtained. The fuzzy comprehensive evaluation unit first determines the set of comments and defines a corresponding membership function for each comment; then, it substitutes the actual observed values of each risk indicator into the corresponding membership function to calculate their respective membership vectors; finally, it performs a weighted synthesis operation based on the pre-determined indicator weight vector and membership matrix to obtain the comprehensive risk membership vector; and selects the comment with the highest membership degree as the final risk level determination result.
10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method described in claim 1.