Enterprise information management system based on artificial intelligence

Abstract

The invention relates to the technical field of enterprise information management, in particular to an enterprise information management system based on artificial intelligence, and is used for solving the problems that an enterprise cannot adjust a strategy in time, resource configuration is difficult to optimize and economic benefits are reduced due to the fact that accurate future sales trend prediction cannot be provided in the prior art. According to the method, the historical operation data is analyzed through the sales trend prediction module, the comprehensive feature vector XQT is generated, and the machine learning model is trained, so that accurate future sales trend prediction is provided, the accuracy of decision making is improved, the flexibility and adaptability of enterprises to deal with market changes are enhanced, and the market competitiveness is improved. The rapid processing and feedback mechanism of the real-time operation data enables enterprises to adjust strategies in time, optimize resource configuration, reduce cost and improve economic benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise information management, and more specifically, to an enterprise information management system based on artificial intelligence. Background Art

[0002] Traditional enterprise information management systems mainly rely on manual classification, sorting, and analysis of data. This method is not only time-consuming and laborious but also prone to errors. With the rapid development of artificial intelligence technology, its application in the field of enterprise information management has become a trend. However, the existing enterprise information management systems based on artificial intelligence in the current market still have deficiencies in terms of intelligence level, data processing efficiency, and decision-making support capabilities.

[0003] The patent application with the publication number CN119106430A discloses a distributed architecture enterprise information management system, including an information system processing unit and an enterprise system monitoring unit. The information system processing unit is used to upgrade and maintain the enterprise management system, so that the enterprise system information is transmitted to the enterprise system monitoring unit. In the present invention, the old data in the system is upgraded and replaced through the information upgrade module, and the data in the old system is imported into the new system. The data management module uniformly manages the data in the new system to improve the integrity of the data. Then, through the system integration module, the public data, hardware, and business processes in the enterprise are integrated to facilitate enterprise business processing and sharing. The data sharing module can reduce duplicate reporting of similar data, enabling enterprise managers to accurately find the data. The maintenance and detection module can sort out the uploaded data stream to achieve partial data sharing and partial data encryption.

[0004] However, the above reference patent realizes the efficient integration and secure management of enterprise data through data upgrade, integration, sharing, and security monitoring modules, improving management efficiency and protecting business secrets. However, it cannot comprehensively consider multiple key indicators, cannot comprehensively depict the sales trend, is not convenient for subsequent analysis and modeling, cannot comprehensively and accurately evaluate the customer churn risk, cannot convert the churn risk into a quantifiable value, is not convenient for analysis and decision-making, and at the same time cannot provide an accurate prediction of future sales trends, making it difficult for enterprises to adjust strategies in a timely manner, optimize resource allocation, and reduce economic benefits.

[0005] Therefore, we propose an enterprise information management system based on artificial intelligence for the above problems. Summary of the Invention

[0006] The object of the present invention is to provide an enterprise information management system based on artificial intelligence, which solves the problems in the prior art that multiple key indicators cannot be comprehensively considered, the sales trend cannot be comprehensively characterized, it is not convenient for subsequent analysis and modeling, the customer churn risk cannot be comprehensively and accurately evaluated, the churn risk cannot be converted into a quantifiable value, it is not convenient for analysis and decision-making, and at the same time, an accurate future sales trend prediction cannot be provided, resulting in the enterprise being unable to adjust its strategies in a timely manner, difficult to optimize resource allocation, and reduced economic benefits.

[0007] The object of the present invention is achieved by the following technical solutions:

[0008] An enterprise information management system based on artificial intelligence, which is applied to an enterprise information management platform, includes:

[0009] A data collection module, which is used to collect the operation data of the target enterprise in real time and perform preprocessing operations on the collected enterprise operation data;

[0010] A feature extraction module, which is used to extract sales trend feature data from the preprocessed operation data, fuse the extracted sales trend feature data, and generate a sales trend comprehensive feature vector;

[0011] A customer churn monitoring module, which is used to monitor the churn evaluation parameters of the target customer in real time and monitor and evaluate the churn risk of the target customer;

[0012] A sales trend prediction module, which is used to collect the historical operation data of the target enterprise, construct a sales trend prediction model, and predict the future sales trend of the target enterprise through the model;

[0013] A decision support module, which is used to generate an optimal decision-making plan according to the customer churn risk monitoring results and the enterprise sales trend prediction results.

[0014] As a preferred implementation manner of the present invention, the specific process of the feature extraction module for extracting sales trend feature data is as follows:

[0015] Obtain the preprocessed operation data of the target enterprise. The operation data includes sales amount, sales volume, sales price, total market sales amount, total production cost price, total number of customers, and number of repeat customers, generate a collection period, and set the collection period duration to one year. Extract sales trend feature data from the preprocessed operation data. The sales trend feature data includes sales amount change rate, sales volume change rate, market share change rate, profit rate, and repurchase rate;

[0016] The specific steps for extracting the sales amount change rate are as follows:

[0017] Obtain the sales amount value S(t) of each product at each time point, and use the difference formula to calculate the sales amount change between adjacent time points:

[0018] ΔS(t) = S(t) - S(t - 1), where S(t) represents the sales amount at time point t, and S(t - 1) represents the sales amount at time point t - 1;

[0019] The average change rate XE(t) of the sales amount at time point t is calculated using the following formula:

[0020] where Δt represents the sampling time interval.

[0021] As a preferred embodiment of the present invention, the specific steps for extracting the sales volume change rate are as follows:

[0022] Obtain the sales volume value Q(t) of each product at each time point t, and use the difference formula to calculate the change in sales volume between adjacent time points:

[0023] ΔQ(t) = Q(t) - Q(t - 1), where Q(t) represents the sales volume at time point t, and Q(t - 1) represents the sales volume at time point t - 1;

[0024] The absolute change rate XL(t) of the sales volume at time point t is calculated using the following formula:

[0025]

[0026] The specific steps for extracting the market share change rate are as follows:

[0027] Obtain the total product sales amount SQ(t) of the target enterprise during the collection period, obtain the total market sales amount SS(t), and calculate the market share at each time point t using the following formula:

[0028] where SQ(t) represents the total product sales amount at time point t, and SS(t) represents the total market sales amount at the same time point t;

[0029] Use the difference formula to calculate the change in market share between adjacent time points:

[0030] ΔM = M(t) - M(t - 1), where M(t) represents the market share at time point t, and M(t - 1) represents the market share at time point t - 1;

[0031] The absolute change rate SF(t) of the market share at time point t is calculated using the following formula:

[0032]

[0033] As a preferred embodiment of the present invention, the specific steps for extracting the profit margin are as follows:

[0034] Obtain the total product sales amount SQ(t) of the target enterprise within the collection period, and obtain the total production cost CZ(t) of the target enterprise within the same time period;

[0035] Use the following formula to calculate the gross profit P(t) at each time point t:

[0036] P(t) = SQ(t) - CZ(t), where SQ(t) represents the total product sales amount at time point t, and CZ(t) represents the total production cost at the same time point t;

[0037] Use the following formula to calculate the profit rate LR(t):

[0038]

[0039] The specific steps to extract the repurchase rate are as follows:

[0040] Obtain the total number of customers of the target enterprise within the collection period, obtain the number of returning customers of the target enterprise within the collection period, and use the following formula to calculate the repurchase rate HG:

[0041]

[0042] Integrate the extracted sales trend characteristic data such as the sales amount change rate, sales volume change rate, market share change rate, profit rate, and repurchase rate to generate a comprehensive sales trend feature vector XQT, which is represented by the following expression:

[0043] XQT = [XE(t), XL(t), SF(t), LR(t), HG].

[0044] As a preferred embodiment of the present invention, the specific process of the customer churn monitoring module for monitoring and evaluating the churn risk of target customers is as follows:

[0045] Obtain the churn evaluation parameters of the target customer. The churn evaluation parameters include the duration since the last purchase, the number of repurchases, the transaction amount, and the repurchase rate. Generate a monitoring period and divide the monitoring period into multiple monitoring time periods;

[0046] Obtain the duration since the last purchase of the target customer in multiple monitoring time periods, and calculate the arithmetic mean of the obtained durations since the last purchase. Denote the arithmetic mean of the multiple durations since the last purchase as the average duration since the last purchase PZS.

[0047] As a preferred embodiment of the present invention, obtain the number of repurchases of the target customer in multiple monitoring time periods, and calculate the arithmetic mean of the obtained number of repurchases. Denote the arithmetic mean of the multiple number of repurchases as the average number of repurchases PHC;

[0048] Obtain the transaction amounts of the target customer in multiple monitoring periods, calculate the arithmetic mean of the obtained multiple transaction amounts, and denote the arithmetic mean of the multiple transaction amounts as the average transaction amount PJJ;

[0049] Obtain the repurchase rates of the target customer in multiple monitoring periods, calculate the arithmetic mean of the obtained multiple repurchase rates, and denote the arithmetic mean of the multiple repurchase rates as the average repurchase rate PHL.

[0050] As a preferred implementation manner of the present invention, obtain the average duration PZS since the last purchase, the average number of repurchases PHC, the average transaction amount PJJ, and the average repurchase rate PHL, and calculate the churn risk assessment coefficient LFP through the following formula:

[0051]

[0052] Where f1, f2, f3, and f4 are all preset proportional factor coefficients, f4 > f3 > f2 > f1 > 0, and compare the churn risk assessment coefficient LFP with the preset churn risk assessment coefficient threshold:

[0053] If the churn risk assessment coefficient LFP is less than the preset churn risk assessment coefficient threshold, it indicates that the possibility of the target customer churning is very small and is in a low-risk state;

[0054] If the churn risk assessment coefficient LFP is greater than or equal to the preset churn risk assessment coefficient threshold, it indicates that the possibility of the target customer churning is relatively large and is in a high-risk state.

[0055] As a preferred implementation manner of the present invention, the specific process of the sales trend prediction module constructing a sales trend prediction model and predicting the future sales trend of the target enterprise through the model is as follows:

[0056] Obtain the historical operation data of the target enterprise, generate a collection period, divide the collection period into multiple collection periods, perform preprocessing operations on the historical operation data, extract sales trend feature data from the historical operation data in multiple collection periods, and fuse the extracted sales trend feature data to generate multiple groups of sales trend comprehensive feature vectors XQT;

[0057] Use the generated sales trend comprehensive feature vector XQT as the input of the machine learning model, and use the change amount of the sales amount in a future period corresponding to each group of sales trend comprehensive feature vectors XQT as the output of the machine learning model. With the change amount of the sales amount in a future period as the prediction target and minimizing the sum of the prediction errors of the training data as the training target, train the machine learning model until the sum of the prediction errors reaches convergence and then stop training to obtain the sales trend prediction model.

[0058] As a preferred embodiment of the present invention, the expression formula of the sales trend prediction model is as follows:

[0059] ΔXE = η1·XE(t) + η2·XL(t) + η3·SF(t) + η4·LR(t) + η5·HG + λ;

[0060] Where ΔXE represents the change in sales amount over a future period of time, η1, η2, η3, η4, and η5 are all regression coefficients, λ is a random error term, ΔS represents the change in sales amount over a future period of time, and ΔQ represents the change in sales volume over a future period of time;

[0061] Obtain real-time operation data, convert it into the corresponding comprehensive feature vector XQT of the sales trend, and input it into the sales trend prediction model. Through the sales trend prediction model, obtain the real-time change in sales amount ΔXE over a future period of time.

[0062] As a preferred embodiment of the present invention, the specific process for the decision support module to generate an optimal decision plan is as follows:

[0063] Obtain the customer churn risk monitoring result and the enterprise sales trend prediction result. The customer churn risk monitoring result is that the customer churn risk status is low risk or high risk, and the enterprise sales trend prediction result is the change in sales volume and the change in sales amount over a future period of time. Generate an optimal decision plan based on the customer churn risk monitoring result and the enterprise sales trend prediction result.

[0064] Compared with the prior art, the advantages of the present invention are as follows:

[0065] (1) In the present invention, through the feature extraction module, multiple key indicators are comprehensively considered, the sales trend is comprehensively characterized, feature extraction is performed based on actual operation data, subjective speculation is avoided, the objectivity and reliability of the results are improved, and the sales trend is converted into quantifiable numerical features, which is convenient for subsequent analysis and modeling;

[0066] (2) In the present invention, through the customer churn monitoring module, the customer churn risk can be monitored in real time, potential problems can be discovered in a timely manner, multiple key indicators are comprehensively considered, the evaluation is more comprehensive and accurate, and the churn risk is converted into quantifiable numerical values, which is convenient for analysis and decision-making;

[0067] (3) In the present invention, through the sales trend prediction module, historical operation data is analyzed, the comprehensive feature vector XQT is generated and the machine learning model is trained to provide accurate future sales trend prediction. This method not only improves the accuracy of decision-making, but also enhances the flexibility and adaptability of the enterprise to market changes. The rapid processing and feedback mechanism of real-time operation data enables the enterprise to adjust strategies in a timely manner, optimize resource allocation, reduce costs, and improve economic benefits. Brief Description of the Drawings

[0068] Figure 1 It is a system block diagram of the first embodiment in the present invention;

[0069] Figure 2 It is a system block diagram of the second embodiment in the present invention;

[0070] Figure 3 It is a schematic diagram of the logic flow in the first embodiment of the present invention. Detailed Description of the Embodiments

[0071] Next, the accompanying drawings in the embodiments of the present invention will be combined; the technical solutions in the embodiments of the present invention will be clearly and completely described; obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0072] Embodiment 1: As Figure 1 and Figure 3 shown, an enterprise information management system based on artificial intelligence proposed by the present invention is applied to an enterprise information management platform, including:

[0073] A data acquisition module, configured to collect the operation data of the target enterprise in real time, and perform preprocessing operations on the collected enterprise operation data. The preprocessing operations include but are not limited to data cleaning, filtering processing, and normalization processing;

[0074] Through preprocessing operations such as data cleaning, filtering, and normalization, the data acquisition module significantly improves the data quality. Data cleaning removes outliers and redundancy to ensure accuracy; filtering processing removes noise and smooths the data; normalization unifies the data scale, accelerates model convergence and improves performance; these operations together optimize the data reliability and provide efficient and accurate support for analysis and decision-making.

[0075] A feature extraction module, configured to extract sales trend feature data from the preprocessed operation data, and fuse the extracted sales trend feature data to generate a sales trend comprehensive feature vector;

[0076] The specific process of the feature extraction module extracting the sales trend feature data is as follows:

[0077] Obtain the pre - processed operation data of the target enterprise. The operation data includes sales amount, sales volume, sales price, total market sales amount, total production cost price, total number of customers, and number of repurchasing customers. Generate a collection period, and set the duration of the collection period to one year. Extract sales trend feature data from the pre - processed operation data. The sales trend feature data includes sales amount change rate, sales volume change rate, market share change rate, profit margin, and repurchase rate;

[0078] The specific steps to extract the sales amount change rate are as follows:

[0079] Obtain the sales amount value S(t) of each product at each time point, and use the difference formula to calculate the sales amount change between adjacent time points:

[0080] ΔS(t) = S(t) - S(t - 1), where S(t) represents the sales amount at time point t, and S(t - 1) represents the sales amount at time point t - 1;

[0081] Use the following formula to calculate the average change rate XE(t) of the sales amount at time point t:

[0082] where Δt represents the sampling time interval;

[0083] The specific steps to extract the sales volume change rate are as follows:

[0084] Obtain the sales volume value Q(t) of each product at each time point t, and use the difference formula to calculate the sales volume change between adjacent time points:

[0085] ΔQ(t) = Q(t) - Q(t - 1), where Q(t) represents the sales volume at time point t, and Q(t - 1) represents the sales volume at time point t - 1;

[0086] Use the following formula to calculate the absolute change rate XL(t) of the sales volume at time point t:

[0087]

[0088] The specific steps to extract the market share change rate are as follows:

[0089] Obtain the total product sales amount SQ(t) of the target enterprise during the collection period, obtain the total market sales amount SS(t), and use the following formula to calculate the market share at each time point t:

[0090] where SQ(t) represents the total product sales amount at time point t, and SS(t) represents the total market sales amount at the same time point t;

[0091] Calculate the change in market share between adjacent time points using the difference formula:

[0092] ΔM = M(t) - M(t - 1), where M(t) represents the market share at time point t and M(t - 1) represents the market share at time point t - 1;

[0093] Calculate the absolute change rate SF(t) of the market share at time point t using the following formula:

[0094]

[0095] The specific steps to extract the profit margin are as follows:

[0096] Obtain the total product sales amount SQ(t) of the target enterprise during the collection period, and obtain the total production cost price CZ(t) of the target enterprise during the same period, which usually includes production costs, raw material costs, direct labor costs, etc.;

[0097] Calculate the gross profit P(t) at each time point t using the following formula:

[0098] P(t) = SQ(t) - CZ(t), where SQ(t) represents the total product sales amount at time point t and CZ(t) represents the total production cost price at the same time point t;

[0099] Calculate the profit margin LR(t) using the following formula:

[0100]

[0101] The specific steps to extract the repurchase rate are as follows:

[0102] Obtain the total number of customers of the target enterprise during the collection period, obtain the number of returning customers of the target enterprise during the collection period, and calculate the repurchase rate HG using the following formula:

[0103]

[0104] Integrate the extracted sales trend characteristic data such as the sales change rate, sales volume change rate, market share change rate, profit margin, and repurchase rate to generate a comprehensive sales trend characteristic vector XQT, which is represented by the following expression:

[0105] XQT = [XE(t), XL(t), SF(t), LR(t), HG];

[0106] The feature extraction module comprehensively considers multiple key indicators, fully depicts the sales trend, extracts features based on actual operation data, avoids subjective speculation, improves the objectivity and reliability of the results, and converts the sales trend into quantifiable numerical features for convenient subsequent analysis and modeling.

[0107] A customer churn monitoring module, which is used to monitor the churn evaluation parameters of target customers in real time and monitor and evaluate the churn risk of target customers;

[0108] The specific process of the customer churn monitoring module for monitoring and evaluating the churn risk of target customers is as follows:

[0109] Obtain the churn evaluation parameters of the target customer. The churn evaluation parameters include the duration since the last purchase, the number of repurchases, the transaction amount, and the repurchase rate. Generate a monitoring period and divide the monitoring period into multiple monitoring time slots;

[0110] Obtain the duration since the last purchase of the target customer in multiple monitoring time slots, and calculate the arithmetic mean of the obtained multiple durations since the last purchase. Denote the arithmetic mean of the multiple durations since the last purchase as the average duration since the last purchase PZS;

[0111] Obtain the number of repurchases of the target customer in multiple monitoring time slots, and calculate the arithmetic mean of the obtained multiple numbers of repurchases. Denote the arithmetic mean of the multiple numbers of repurchases as the average number of repurchases PHC;

[0112] Obtain the transaction amount of the target customer in multiple monitoring time slots, and calculate the arithmetic mean of the obtained multiple transaction amounts. Denote the arithmetic mean of the multiple transaction amounts as the average transaction amount PJJ;

[0113] Obtain the repurchase rate of the target customer in multiple monitoring time slots, and calculate the arithmetic mean of the obtained multiple repurchase rates. Denote the arithmetic mean of the multiple repurchase rates as the average repurchase rate PHL;

[0114] After obtaining the average duration since the last purchase PZS, the average number of repurchases PHC, the average transaction amount PJJ, and the average repurchase rate PHL, calculate the churn risk assessment coefficient LFP through the following formula:

[0115]

[0116] Where f1, f2, f3, and f4 are all preset proportional factor coefficients, f4 > f3 > f2 > f1 > 0. Compare the churn risk assessment coefficient LFP with the preset churn risk assessment coefficient threshold:

[0117] If the churn risk assessment coefficient LFP is less than the preset churn risk assessment coefficient threshold, it indicates that the possibility of the target customer churning is very small and is in a low-risk state;

[0118] If the churn risk assessment coefficient LFP is greater than or equal to the preset churn risk assessment coefficient threshold, it indicates that the possibility of the target customer churning is relatively large and is in a high-risk state;

[0119] The customer churn monitoring module can monitor the customer churn risk in real time, discover potential problems in a timely manner, comprehensively consider multiple key indicators, evaluate more comprehensively and accurately, convert the churn risk into a quantifiable value, which is convenient for analysis and decision-making.

[0120] The sales trend prediction module is used to collect the historical operation data of the target enterprise, construct a sales trend prediction model, and predict the future sales trend of the target enterprise through the model;

[0121] The specific process of the sales trend prediction module constructing a sales trend prediction model and predicting the future sales trend of the target enterprise through the model is as follows:

[0122] Obtain the historical operation data of the target enterprise, generate a collection period, divide the collection period into multiple collection time periods, perform preprocessing operations on the historical operation data, extract sales trend feature data from the historical operation data within multiple collection time periods, fuse the extracted sales trend feature data, and generate multiple groups of sales trend comprehensive feature vectors XQT;

[0123] Use the generated sales trend comprehensive feature vector XQT as the input of the machine learning model, and use the sales volume change amount in the future period corresponding to each group of sales trend comprehensive feature vectors XQT as the output of the machine learning model. With the sales volume change amount in the future period as the prediction target and minimizing the sum of prediction errors of the training data as the training target, train the machine learning model until the sum of prediction errors reaches convergence and then stop training to obtain the sales trend prediction model;

[0124] The expression formula of the sales trend prediction model is as follows:

[0125] ΔXE = η1·XE(t) + η2·XL(t) + η3·SF(t) + η4·LR(t) + η5·HG + λ;

[0126] Among them, ΔXE represents the sales volume change amount in the future period, η1, η2, η3, η4, and η5 are all regression coefficients, and λ is a random error term;

[0127] Obtain the real-time operation data, convert it into the corresponding sales trend comprehensive feature vector XQT and input it into the sales trend prediction model, and obtain the real-time sales volume change amount ΔXE in the future period through the sales trend prediction model;

[0128] By analyzing historical operation data through the sales trend prediction module, a comprehensive feature vector XQT is generated and a machine learning model is trained to provide accurate future sales trend predictions. This method not only improves the accuracy of decision-making but also enhances the flexibility and adaptability of enterprises in coping with market changes. The rapid processing and feedback mechanism of real-time operation data enable enterprises to adjust strategies in a timely manner, optimize resource allocation, reduce costs, and improve economic efficiency.

[0129] Embodiment 2: The technical solution of this embodiment of the present invention is different from that of Embodiment 1 in that:

[0130] As Figure 2 shown, a decision support module is used to generate an optimal decision-making plan according to the customer churn risk monitoring result and the enterprise sales trend prediction result;

[0131] The specific process of the decision support module generating the optimal decision-making plan is as follows:

[0132] Obtain the customer churn risk monitoring result and the enterprise sales trend prediction result. The customer churn risk monitoring result is that the customer churn risk status is low risk or high risk, and the enterprise sales trend prediction result is the change in sales volume in a future period. According to the customer churn risk monitoring result and the enterprise sales trend prediction result, an optimal decision-making plan is generated. The specific content of the optimal decision-making plan is:

[0133] If the customer churn risk status is low risk and the predicted sales volume increases, maintain the existing strategy, increase market promotion efforts, develop new products or services, increase advertising and promotion activities, and conduct market research and new product R & D;

[0134] If the customer churn risk status is low risk and the predicted sales volume decreases, it is necessary to analyze the reasons for the sales decline, adjust the product or price strategy, conduct market research, competitor analysis, and improve customer service and product price or function;

[0135] If the customer churn risk status is high risk and the predicted sales volume increases, retention measures should be taken for high-risk customer groups. Identify high-risk customers through the customer relationship management system, provide personalized services, and carry out customer loyalty programs;

[0136] If the customer churn risk status is high risk and the predicted sales volume decreases, emergency measures need to be taken, such as providing discount offers to retain customers, while conducting market research, competitor analysis, and adjusting the product strategy or price strategy;

[0137] Through the decision-making support module, enterprises can make scientific and accurate decisions based on objective data, reducing subjective assumptions. This module responds in real time to market changes and customer feedback, provides personalized solutions, optimizes resource allocation, and reduces the churn rate of high-risk customers. The automated process reduces manual intervention, improves efficiency, and can expand rules and metrics as needed, enhancing the complexity and refinement of decision-making, thereby strengthening the enterprise's market adaptability and overall operational efficiency.

[0138] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its improved concept, making equivalent substitutions or changes, should be covered by the protection scope of the present invention.

Claims

1. An enterprise information management system based on artificial intelligence, applied to an enterprise information management platform, characterized in that: include: The data collection module is used to collect the target enterprise's operating data in real time and perform pre-processing operations on the collected enterprise operating data; A feature extraction module is used to extract sales trend feature data from the preprocessed operation data, fuse the extracted sales trend feature data, and generate a sales trend comprehensive feature vector; The customer churn monitoring module is used to monitor the churn assessment parameters of target customers in real time and monitor and assess the churn risk of target customers; The sales trend forecasting module is used to collect the historical operating data of the target enterprise, build a sales trend forecasting model, and use the model to forecast the future sales trend of the target enterprise; The decision support module is used to generate the optimal decision plan based on the customer churn risk monitoring results and the enterprise sales trend forecast results.

2. The enterprise information management system based on artificial intelligence according to claim 1, characterized in that: The specific process of extracting sales trend feature data by the feature extraction module is as follows: Obtain the pre-processed operating data of the target enterprise, including sales, sales volume, sales price, total market sales, total production cost, total number of customers and number of repurchased customers, generate a collection cycle, and set the collection cycle duration to one year. Extract sales trend feature data from the pre-processed operating data, including sales change rate, sales volume change rate, market share change rate, profit margin and repurchase rate; The specific steps to extract the sales change rate are as follows: Get the sales value S(t) of each product at each time point, and use the difference formula to calculate the change in sales between adjacent time points: ΔS(t)=S(t)-S(t-1), where S(t) represents the sales at time point t, and S(t-1) represents the sales at time point t-1; Use the following formula to calculate the average rate of change of sales at time point t, XE(t): Where Δt represents the sampling time interval.

3. The enterprise information management system based on artificial intelligence according to claim 2 is characterized in that: The specific steps to extract the sales volume change rate are as follows: Get the sales volume value Q(t) of each product at each time point t, and use the difference formula to calculate the sales volume change between adjacent time points: ΔQ(t)=Q(t)-Q(t-1), where Q(t) represents the sales volume at time point t, and Q(t-1) represents the sales volume at time point t-1; The absolute rate of change in sales volume at time point t, XL(t), is calculated using the following formula: The specific steps for extracting the market share change rate are as follows: Obtain the total product sales SQ(t) of the target enterprise during the collection period, obtain the total sales SS(t) of the entire market, and use the following formula to calculate the market share at each time point t: Where SQ(t) represents the total product sales at time point t, and SS(t) represents the total market sales at the same time point t; Use the difference formula to calculate the change in market share between adjacent time points: ΔM=M(t)-M(t-1), where M(t) represents the market share at time point t, and M(t-1) represents the market share at time point t-1; The absolute rate of change in market share at time point t is calculated using the following formula:

4. The enterprise information management system based on artificial intelligence according to claim 3 is characterized in that: The specific steps to extract the profit rate are as follows: Obtain the total product sales SQ(t) of the target enterprise during the collection period, and obtain the total production cost CZ(t) of the target enterprise during the same period; The gross profit P(t) at each time point t is calculated using the following formula: P(t) = SQ(t) - CZ(t), where SQ(t) represents the total sales of products at time point t, and CZ(t) represents the total production cost at the same time point t; The profit margin LR(t) is calculated using the following formula: The specific steps for extracting the repurchase rate are as follows: Obtain the total number of customers of the target enterprise during the collection period, obtain the number of returning customers of the target enterprise during the collection period, and use the following formula to calculate the repurchase rate HG: The extracted sales trend feature data such as sales change rate, sales volume change rate, market share change rate, profit margin and repurchase rate are integrated to generate a sales trend comprehensive feature vector XQT, which is expressed by the following expression: XQT=[XE(t),XL(t),SF(t),LR(t),HG].

5. The enterprise information management system based on artificial intelligence according to claim 1, characterized in that: The specific process of the customer churn monitoring module monitoring and evaluating the churn risk of target customers is as follows: Obtain the target customer's churn assessment parameters, including the time since the last purchase, the number of repurchases, the transaction amount, and the repurchase rate, generate a monitoring cycle, and divide the monitoring cycle into multiple monitoring periods; The time from the target customer to the last purchase in multiple monitoring periods is obtained, and the arithmetic mean of the multiple time from the last purchase is calculated, and the arithmetic mean of the multiple time from the last purchase is recorded as the average time from the last purchase PZS.

6. The enterprise information management system based on artificial intelligence according to claim 5, characterized in that: Obtain the number of repurchases of the target customer in multiple monitoring periods, calculate the arithmetic average of the multiple repurchase numbers, and record the arithmetic average of the multiple repurchase numbers as the average repurchase number PHC; Obtain the transaction amount of the target customer in multiple monitoring periods, and calculate the arithmetic average of the multiple transaction amounts obtained, and record the arithmetic average of the multiple transaction amounts as the average transaction amount PJJ; The repurchase rate of the target customers in multiple monitoring periods is obtained, and the arithmetic mean of the multiple repurchase rates is calculated, and the arithmetic mean of the multiple repurchase rates is recorded as the average repurchase rate PHL.

7. The enterprise information management system based on artificial intelligence according to claim 6 is characterized in that: Obtain the average time from the last purchase PZS, the average number of repurchases PHC, the average transaction amount PJJ, and the average repurchase rate PHL, and calculate the churn risk assessment coefficient LFP using the following formula: Among them, f1, f2, f3 and f4 are all preset proportional factor coefficients, f4>f3>f2>f1>0, and the churn risk assessment coefficient LFP is compared with the preset churn risk assessment coefficient threshold: If the churn risk assessment coefficient LFP is less than the preset churn risk assessment coefficient threshold, it means that the target customer is unlikely to churn and is in a low risk state; If the churn risk assessment coefficient LFP is greater than or equal to the preset churn risk assessment coefficient threshold, it indicates that the target customer is more likely to churn and is in a high-risk state.

8. The enterprise information management system based on artificial intelligence according to claim 1, characterized in that: The specific process of the sales trend prediction module constructing a sales trend prediction model and predicting the future sales trend of the target enterprise through the model is as follows: Obtain the historical operation data of the target enterprise, generate a collection cycle, divide the collection cycle into multiple collection periods, perform preprocessing operations on the historical operation data, extract sales trend feature data from the historical operation data in multiple collection periods, fuse the extracted sales trend feature data, and generate multiple sets of sales trend comprehensive feature vectors XQT; The generated sales trend comprehensive feature vector XQT is used as the input of the machine learning model, and the sales change in the future period corresponding to each group of sales trend comprehensive feature vectors XQT is used as the output of the machine learning model. The sales change in the future period is taken as the prediction target, and minimizing the sum of prediction errors of training data is taken as the training target. The machine learning model is trained until the sum of prediction errors converges and the training is stopped to obtain a sales trend prediction model.

9. The enterprise information management system based on artificial intelligence according to claim 8, characterized in that: The sales trend forecasting model is expressed as follows: ΔXE=eta1· Among them, ΔXE represents the change in sales in the future, η1, η2, η3, η4 and η5 are all regression coefficients, λ is the random error term, ΔS represents the change in sales in the future, and ΔQ represents the change in sales volume in the future; Real-time operation data is obtained, converted into the corresponding sales trend comprehensive feature vector XQT, and input into the sales trend prediction model. The real-time sales change ΔXE in the future period is obtained through the sales trend prediction model.

10. The enterprise information management system based on artificial intelligence according to claim 1, characterized in that: The specific process of the decision support module generating the optimal decision solution is as follows: Obtain the customer churn risk monitoring results and enterprise sales trend forecasting results. The customer churn risk monitoring results indicate whether the customer churn risk status is low risk or high risk. The enterprise sales trend forecasting results indicate the change in sales volume and sales revenue in the future. Generate an optimal decision plan based on the customer churn risk monitoring results and enterprise sales trend forecasting results.

Images