Enterprise data asset auditing method and system based on big data

Abstract

The invention belongs to the technical field of data asset management, and discloses an enterprise data asset auditing method and system based on big data, and the method comprises the steps: collecting financial information data, business operation data, management decision data and external industry data of an enterprise; preprocessing the collected financial information data, business operation data, management decision data and external industry data to obtain a financial feature data set, a business operation feature data set, a management decision feature data set and an external industry feature data set; performing feature extraction and weighted fusion on the financial feature data set, the business operation feature data set, the management decision feature data set and the external industry feature data set to obtain a comprehensive feature data set; constructing an audit risk prediction model, inputting the comprehensive feature data set into the audit risk prediction model, and predicting to obtain an audit risk coefficient of the enterprise data assets; and a more comprehensive view angle is provided, and the data asset risk of the enterprise can be evaluated more accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of data asset management, and more specifically, to a method and system for enterprise data asset auditing based on big data. Background Art

[0002] The patent application with the publication number CN115375417A discloses a comprehensive financial audit system based on big data, including a data collection module, a data analysis module, a learning prediction module and a database. The data collection module collects structured and semi-structured enterprise intangible asset evaluation data and data information of related industries based on big data technology, and collects enterprise intangible asset audit data at the same time. The data analysis module analyzes and compares the data collected by the data collection module with the historical data in the database to conduct a comprehensive analysis of the audited data of the enterprise's intangible assets. This comprehensive financial audit system based on big data improves the traditional asset evaluation, combines big data technology to improve the reliability of the evaluation results of intangible assets, and at the same time improves the audit quality and reduces the risk of evaluation differences.

[0003] The existing enterprise data asset audit methods and systems have the following main problems: Failure to use cross-validation to evaluate the performance of the model on different data subsets may increase the risk of overfitting or underfitting, affecting the overall performance and reliability of the model; lack of cross-validation means that the performance evaluation of the model depends on the partitioning results of a single data set; failure to consider the fold limit formula to reasonably adjust the fold of cross-validation, which may be very time-consuming and waste computing resources for situations with many hyperparameter combinations; too many computing resources are consumed on unnecessary model training and validation, reducing computing efficiency; lack of standardized hyperparameter optimization processes such as grid search and cross-validation, and the results of different experiments may vary significantly due to different hyperparameter settings; Without using dynamically adjusted threshold formulas, risk assessment usually relies on fixed threshold settings. This fixed threshold may not be able to adapt to the risk characteristics of different enterprises and industries, resulting in a lack of flexibility in the assessment results and an inability to accurately reflect the actual risks faced by the enterprise; without considering the use of statistical features such as median, standard deviation, and percentile to adjust the threshold, the assessment results are easily affected by outliers or extreme data; outliers may cause the fixed threshold to be too high or too low, affecting the accuracy and stability of risk assessment.

[0004] Instead of optimizing the determination of risk thresholds through coefficient adjustment models, traditional methods rely on simple statistics or artificially set fixed values. This results in inaccurate risk thresholds that cannot accurately reflect the actual risk status of the enterprise. Failure to optimize the adjustment coefficients using factors such as the number of features and variance will reduce the robustness and flexibility of the evaluation model. The model may be overly sensitive or not sensitive enough to certain specific data features, making it difficult to cope with complex and changing risk scenarios, resulting in unstable evaluation results.

[0005] In view of this, the present invention proposes an enterprise data asset auditing method and system based on big data to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for auditing enterprise data assets based on big data, comprising: S1. Collect the company's financial information data, business operation data, management decision data and external industry data; S2. Preprocess the collected financial information data, business operation data, management decision data and external industry data to obtain a financial feature data set, a business operation feature data set, a management decision feature data set and an external industry feature data set; Extract and weightedly fuse the financial feature data set, business operation feature data set, management decision feature data set, and external industry feature data set to obtain a comprehensive feature data set; S3. Construct an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; S4. Compare the predicted audit risk coefficient of the enterprise data assets with the preset audit risk coefficient threshold of the enterprise data assets to evaluate the audit risk level faced by the enterprise data assets; S5. Develop corresponding asset optimization management strategies based on the audit risk level faced by enterprise data assets.

[0007] Furthermore, the collected enterprise's financial information data includes: income and expenditure data, asset data, liability data, cash flow data and tax data; business operation data includes production data, inventory data, sales data, supply chain data, customer behavior data and operation performance data; Management decision data include strategic planning data, performance appraisal data, internal audit data, budget and forecast data and decision support data; external industry data include industry market data, economic environment data, policy and regulation data and technology trend data.

[0008] Furthermore, the method of preprocessing the collected financial information data, business operation data, management decision data and external industry data to obtain the financial information data set, business operation data set, management decision data set and external industry data set includes: Use linear interpolation to fill missing values ​​in financial information data, business operation data, management decision data, and external industry data; use the linear relationship between the previous and next non-missing values ​​to calculate the estimated value of the missing value and replace the original missing value; check the uniqueness of financial information data, business operation data, management decision data, and external industry data, and remove duplicate records; Financial information data, business operation data, management decision data and external industry data are standardized and normalized, and data of different units and dimensions are converted into standard normal distribution to obtain financial information data set, business operation data set, management decision data set and external industry data set.

[0009] Furthermore, the method of extracting features and weighted fusion of the financial information dataset, the business operation dataset, the management decision dataset and the external industry dataset to obtain a comprehensive feature dataset includes: The method for extracting features from a financial information data set is: calculating financial indicators and performing financial ratio analysis on the financial information data; performing time series analysis on the acquired financial indicators and extracting the trend, seasonality and cyclical features of the time series data; integrating the acquired financial indicators with the trend, seasonality and cyclical features of the data to obtain a financial feature data set; The method of feature extraction for business operation data set is: perform time series analysis on sales data in business operation data set to extract seasonal, trend and periodic features; perform production efficiency analysis on production data to extract productivity, equipment utilization and production cycle time features; calculate inventory turnover rate, inventory holding cost and supply and demand balance features for inventory data; perform customer feature analysis based on customer behavior data. The specific method is as follows: S41, extracting the RFM characteristics of the customer; calculating the R, F, and M indicators for each customer respectively, and scoring the R, F, and M indicators according to the calculation results; S42. Based on the R, F, and M indicator scores, the customer's scores in the three dimensions of R, F, and M are combined to obtain R, F, and M score combinations, each of which represents a different customer type, and a comprehensive score for each customer is calculated; The comprehensive score calculation formula is: ; in, For the Total score of customers; For the R-index score of each customer; For the F-index score of each customer; For the M-index score of each customer; For the R index score weight coefficient for each customer; For the The F-index score weight coefficient of each customer; For the The M indicator score weight coefficient for each customer; Index for customers; S43, preliminarily grouping the customers according to different score combinations, where each combination represents a different customer type, and restricting the score combinations through a score combination restriction model; The score combination restriction model is: ; in, is the proportion of a certain customer type in the total customer base; is the minimum comprehensive score among the customer groups; is the maximum comprehensive score among the customer groups; is the ratio of the scoring dimension to the total number of customers; is the total number of customers; Types of customers; is the number of scoring dimensions of RFM; S44, standardize the R, F, and M index scores of each customer; the standardization formula is: ; in, Score the standardized indicators; For customers R, F, M index scores; is the mean of the indicator scores; the standard deviation for scoring the indicator; Indexing for customers; The K-means clustering algorithm is used to further segment customers; the method of extracting features from the management decision data set is: SWOT analysis and key performance indicator analysis of the management decision data; The method of feature extraction for external industry data sets is: using time series analysis methods to analyze market trends and extract macroeconomic features; Through the above steps, we obtain the financial information data set, business operation data set, management decision data set and external industry data set after feature extraction; we perform weighted fusion on the financial information data set, business operation data set, management decision data set and external industry data set after feature extraction to obtain a comprehensive feature data set.

[0010] Furthermore, the method of performing weighted fusion on the feature-extracted financial feature data set, business operation feature data set, management decision feature data set and external industry feature data set to obtain a comprehensive feature data set includes: The financial feature dataset, business operation feature dataset, management decision feature dataset and external industry feature dataset after feature extraction are fused through a weighted model to form a comprehensive feature dataset. The financial feature dataset is recorded as , the business operation feature dataset is recorded as , the management decision feature dataset is recorded as ; The external industry characteristic data set is recorded as ; The weighted model is: ; in, is the weight coefficient of the financial feature data set; is the weight coefficient of the business operation feature data set; is the weight coefficient of the management decision feature data set; is the weight coefficient of the external industry characteristic data set.

[0011] Furthermore, the training method of the audit risk prediction model includes: The data set is divided into a training set, a validation set and a test set. The data set includes a historical comprehensive feature data set and the audit risk coefficient of the corresponding enterprise asset data. An audit risk prediction model is constructed. The model includes an input layer, a GRU layer, a fully connected layer and an output layer. The input layer of the audit risk prediction model is used to input the historical comprehensive feature data set. The number of neurons in the input layer matches the number of features in the historical comprehensive feature data set. The GRU layer is used to process the historical comprehensive feature data set, and the number of GRU layers and neurons are adjusted according to the complexity of the task; the fully connected layer is used to provide additional nonlinear transformation; the model output layer is used to output the network signal performance index, and the predicted value is output through a neuron, and the identity function is used as the activation function; the audit risk prediction model is a gated recurrent unit model; The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: ; in, is the mean absolute error; is the number of data sets; For the The true value of the audit risk coefficient of the enterprise asset data; For the The predicted value of the audit risk coefficient of the asset data of an enterprise; Use the training set data to train the model and minimize the loss function using the Adam optimizer. Use the validation set to evaluate the performance of the model and tune the model's hyperparameters until the model performance no longer improves significantly or reaches the preset stopping condition. Use the test set to evaluate the performance of the model in the prediction task, input the current comprehensive feature data set into the trained audit risk prediction model, and obtain the audit risk coefficient of the enterprise data assets.

[0012] Furthermore, the method for tuning the hyperparameters of the model includes: S71, there are two preset hyperparameters: learning rate and the number of GRU layers , use grid search to traverse these two hyperparameter combinations, preset learning rate The value range of is: ; in, is the learning rate The number of values; the preset number of GRU layers The value range of is: ; in, is the number of GRU layers The number of values ​​of ; S72. For each hyperparameter combination, The fold cross validation error formula calculates the error of this hyperparameter combination. fold cross validation error, The fold cross validation error formula is: ; in, For the hyperparameter combination Next, through The average validation error obtained by fold cross validation; For the The value of the learning rate; For the The value of the number of GRU layers; for The number of folds in the cross-validation; For a traversal of all The folded loop vector; For From 1 to The sum of For the Compromise using hyperparameter combinations Validation error after training the model; S73, through Fold restriction formula Number of folds in cross validation To restrict The fold limit formula is: ; in, is the total number of hyperparameter combinations; is the number of hyperparameter types; is the floor rounding function; S74, select The hyperparameter combination with the smallest cross-validation error is taken as the optimal hyperparameter combination of the model.

[0013] Furthermore, the method of comparing the predicted audit risk coefficient of the enterprise data assets with the preset enterprise data asset audit risk coefficient threshold to evaluate the audit risk level faced by the enterprise data assets includes: S81. Preset the audit risk coefficient threshold and limit the audit risk coefficient threshold through the threshold adjustment formula; the threshold adjustment formula is: ; in, is the audit risk factor threshold; is the median of historical audit risk factors; is the standard deviation of the historical audit risk coefficient; is the percentile of the historical audit risk factor; The adjustment factor for the difference between the median and percentile of the historical audit risk factor; S82. Adjustment coefficient for the difference between the median and percentile of historical audit risk coefficients through the coefficient adjustment model Adjustment is made, and the coefficient adjustment model is: ; in, is the number of features that affect the historical audit risk factor; is the number of historical audit risk factors; is the variance of the historical audit risk coefficient; S83, when When the value of is the first quartile, the audit risk coefficient threshold is recorded as ;when When the value of is the third quartile, the audit risk coefficient threshold is recorded as ; S84. According to the preset audit risk coefficient threshold, the audit risk level threshold faced by the enterprise data assets is further divided; the predicted audit risk coefficient is recorded as ; S85. Compare the predicted audit risk coefficient with the preset audit risk coefficient threshold to determine the audit risk level faced by the enterprise data assets; if , then the audit risk level faced by the enterprise data assets is judged to be low risk; if , then the audit risk level faced by the enterprise data assets is determined to be medium risk; if , then the audit risk level faced by the enterprise’s data assets is determined to be high risk.

[0014] Furthermore, the method of formulating a corresponding asset optimization management strategy according to the audit risk level faced by the enterprise data assets includes: For enterprise data assets with low audit risk, the management strategies adopted include data quality inspection, data classification, hierarchical management, setting local data access control permissions and starting local data audit tracking mechanism; For enterprise data assets with moderate audit risk, the management strategies adopted include implementing joint data protection measures, strengthening data access control strategies, conducting comprehensive authority management in combination with role and attribute control, establishing data backup and recovery mechanisms, encrypting data storage and transmission, and implementing user behavior analysis and auditing, detecting abnormal data operations, and generating early warning information through enterprise intelligent management terminals; For enterprise data assets with high audit risks, we adopt a comprehensive optimization management strategy, conduct a comprehensive audit of all data, monitor all data operations, and generate audit reports through enterprise intelligent management terminals. We deploy DLP data leakage prevention technology, use high-strength encryption algorithms to encrypt data, and issue interception instructions through enterprise intelligent management terminals to automatically intercept unauthorized data transmission.

[0015] An enterprise data asset audit system based on big data, comprising: Data collection module, used to collect the company's financial information data, business operation data, management decision data and external industry data; The data processing module is used to pre-process the collected financial information data, business operation data, management decision data and external industry data to obtain financial information data set, business operation data set, management decision data set and external industry data set; Perform feature extraction and weighted fusion on financial information datasets, business operation datasets, management decision datasets, and external industry datasets to obtain a comprehensive feature dataset; The risk prediction module is used to build an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; The risk assessment module is used to compare the predicted audit risk coefficient of enterprise data assets with the preset enterprise data asset audit risk coefficient threshold to assess the audit risk level faced by enterprise data assets; The adjustment and optimization module is used to formulate corresponding optimized data management strategies based on the audit risk level faced by enterprise data assets.

[0016] The technical effects and advantages of the enterprise data asset audit method and system based on big data of the present invention are as follows: Through cross-validation, the performance of the model on different data subsets can be effectively evaluated to ensure the stability of the model training process; multi-fold validation can reduce the impact of random data division on the model evaluation results and improve the reliability of the results; the fold limit formula adjusts the fold according to the number of hyperparameter combinations to avoid unnecessary computational overhead; making model training more efficient, especially when there are many hyperparameter combinations, it can significantly reduce the computational cost; the standardized process of grid search and cross-validation methods makes the experimental results more reproducible; By introducing the threshold adjustment formula and combining the median, standard deviation, percentile and other statistical characteristics of the historical audit risk coefficient, the audit risk coefficient threshold can be adjusted dynamically; it can better adapt to the risk characteristics of different enterprises and industries, avoid the inflexibility caused by fixed thresholds, and make risk assessment more accurate and adaptable; using multiple statistical characteristics such as median, standard deviation, percentile, etc. to calculate the threshold of the audit risk coefficient makes the assessment method more scientific and comprehensive; it can reduce the impact of outliers on the results and provide more stable risk assessment results; The difference between the median and percentile of the historical audit risk coefficient is adjusted through the coefficient adjustment model, and the adjustment coefficient is further optimized using factors such as the number of features and variance. This can enhance the robustness and flexibility of the model, making the determination of risk thresholds more accurate. By using mathematical models and data analysis methods to determine the audit risk level, the subjectivity of the assessment process is reduced. This makes the risk assessment process more standardized, transparent and explainable, which helps to improve the objectivity and consistency of audit and management decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of an enterprise data asset audit method based on big data according to the present invention; Figure 2This is a schematic diagram of the structure of an enterprise data asset audit system based on big data of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1 See also Figure 1 As shown, this embodiment provides an enterprise data asset audit method based on big data, including: S1. Collect the company's financial information data, business operation data, management decision data and external industry data; S2. Preprocess the collected financial information data, business operation data, management decision data and external industry data to obtain a financial feature data set, a business operation feature data set, a management decision feature data set and an external industry feature data set; Extract and weightedly fuse the financial feature data set, business operation feature data set, management decision feature data set, and external industry feature data set to obtain a comprehensive feature data set; S3. Construct an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; S4. Compare the predicted audit risk coefficient of the enterprise data assets with the preset audit risk coefficient threshold of the enterprise data assets to evaluate the audit risk level faced by the enterprise data assets; S5. Develop corresponding asset optimization management strategies based on the audit risk level faced by enterprise data assets.

[0020] The collected financial information data of the enterprise include: income and expenditure data, asset data, liability data, cash flow data and tax data; the business operation data include production data, inventory data, sales data, supply chain data, customer behavior data and operation performance data; Management decision-making data includes strategic planning data, performance appraisal data, internal audit data, budget and forecast data, and decision support data; external industry data includes industry market data, economic environment data, policy and regulation data, and technology trend data; The collected financial information data, business operation data, management decision data and external industry data are collected through the enterprise's internal systems, data integration platforms and external data service platforms.

[0021] The method of preprocessing the collected financial information data, business operation data, management decision data and external industry data to obtain the financial information data set, business operation data set, management decision data set and external industry data set includes: Use linear interpolation to fill missing values ​​in financial information data, business operation data, management decision data, and external industry data; use the linear relationship between the previous and next non-missing values ​​to calculate the estimated value of the missing value and replace the original missing value; check the uniqueness of financial information data, business operation data, management decision data, and external industry data, and remove duplicate records; Financial information data, business operation data, management decision data and external industry data are standardized and normalized, and data of different units and dimensions are converted into standard normal distribution to obtain financial information data set, business operation data set, management decision data set and external industry data set.

[0022] Methods for extracting features and weighted fusion of financial information datasets, business operation datasets, management decision datasets, and external industry datasets to obtain a comprehensive feature dataset include: The method for extracting features from a financial information data set is: calculating financial indicators and performing financial ratio analysis on the financial information data; performing time series analysis on the acquired financial indicators and extracting the trend, seasonality and cyclical features of the time series data; integrating the acquired financial indicators with the trend, seasonality and cyclical features of the data to obtain a financial feature data set; The method of feature extraction for business operation data set is: perform time series analysis on sales data in business operation data set to extract seasonal, trend and periodic features; perform production efficiency analysis on production data to extract productivity, equipment utilization and production cycle time features; calculate inventory turnover rate, inventory holding cost and supply and demand balance features for inventory data; perform customer feature analysis based on customer behavior data. The specific method is as follows: S41. Extract the RFM features of customers; calculate the R, F, and M indicators for each customer respectively, and score the R, F, and M indicators according to the calculation results; calculate the RFM value for each customer: R: calculate the time interval between each customer's most recent purchase date p and the analysis date p'; ​​it can be in days, weeks, or months; F: calculate the total number of purchases for each customer within the specified time window t; M: calculate the total purchase amount for each customer within the specified time window t.

[0023] The values ​​of R, F, and M are divided into multiple levels (usually 1 to 5), and the scoring criteria are as follows: The smaller the R value, the more active the customer is and the higher the score is; the larger the F value, the more frequent the customer purchases and the higher the score is; the larger the M value, the more the customer spends and the higher the score is; S42. Based on the R, F, and M indicator scores, the customer's scores in the three dimensions of R, F, and M are combined to obtain R, F, and M score combinations, each of which represents a different customer type, and a comprehensive score for each customer is calculated; The comprehensive score calculation formula is: ; in, For the Total score of customers; For the R-index score of each customer; For the F-index score of each customer; For the M-index score of each customer; For the R index score weight coefficient for each customer; For the The F-index score weight coefficient of each customer; For the The M indicator score weight coefficient for each customer; Index for customers; S43, preliminarily grouping the customers according to different score combinations, where each combination represents a different customer type, and restricting the score combinations through a score combination restriction model; The score combination restriction model is: ; in, is the proportion of a certain customer type in the total customer base; is the minimum comprehensive score among the customer groups; is the maximum comprehensive score among the customer groups; is the ratio of the scoring dimension to the total number of customers; is the total number of customers; Types of customers; is the number of scoring dimensions of RFM; when When it increases, it means that customers are divided into smaller groups, which can be adjusted to control for different types of score differences.

[0024] If you need to reduce the number of types, you can appropriately reduce the impact of the scoring dimensions on customers, that is, reduce Or increase the total number of customers .

[0025] S44. Standardize the R, F, and M index scores of each customer; the standardization formula is: ; in, Score the standardized indicators; For customers R, F, M index scores; is the mean of the indicator scores; the standard deviation for scoring the indicator; Indexing for customers; The K-means clustering algorithm is used to further segment customers; the method of extracting features from the management decision data set is: SWOT analysis and key performance indicator analysis of the management decision data; The method of feature extraction for external industry data sets is: using time series analysis methods to analyze market trends and extract macroeconomic features; Through the above steps, we obtain the financial information data set, business operation data set, management decision data set and external industry data set after feature extraction; we perform weighted fusion on the financial information data set, business operation data set, management decision data set and external industry data set after feature extraction to obtain a comprehensive feature data set.

[0026] Methods for weighting and fusing the extracted financial feature data set, business operation feature data set, management decision feature data set and external industry feature data set to obtain a comprehensive feature data set include: The financial feature dataset, business operation feature dataset, management decision feature dataset and external industry feature dataset after feature extraction are fused through a weighted model to form a comprehensive feature dataset. The financial feature dataset is recorded as , the business operation feature dataset is recorded as , the management decision feature dataset is recorded as ; The external industry characteristic data set is recorded as ; The weighted model is: ; in, is the weight coefficient of the financial feature data set; is the weight coefficient of the business operation feature data set; is the weight coefficient of the management decision feature data set; is the weight coefficient of the external industry characteristic data set.

[0027] The training methods of audit risk prediction models include; The data set is divided into a training set, a validation set and a test set. The data set includes a historical comprehensive feature data set and the audit risk coefficient of the corresponding enterprise asset data. An audit risk prediction model is constructed. The model includes an input layer, a GRU layer, a fully connected layer and an output layer. The input layer of the audit risk prediction model is used to input the historical comprehensive feature data set. The number of neurons in the input layer matches the number of features in the historical comprehensive feature data set. The GRU layer is used to process the historical comprehensive feature data set, and the number of GRU layers and neurons are adjusted according to the complexity of the task; the fully connected layer is used to provide additional nonlinear transformation; the model output layer is used to output the network signal performance index, and the predicted value is output through a neuron, and the identity function is used as the activation function; the audit risk prediction model is a gated recurrent unit model; The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: ; in, is the mean absolute error; is the number of data sets; For the The true value of the audit risk coefficient of the enterprise asset data; For the The predicted value of the audit risk coefficient of the asset data of an enterprise; Use the training set data to train the model and minimize the loss function using the Adam optimizer. Use the validation set to evaluate the performance of the model and tune the model's hyperparameters until the model performance no longer improves significantly or reaches the preset stopping condition. Use the test set to evaluate the performance of the model in the prediction task, input the current comprehensive feature data set into the trained audit risk prediction model, and obtain the audit risk coefficient of the enterprise data assets.

[0028] Methods for tuning the model's hyperparameters include: S71, there are two preset hyperparameters: learning rate and the number of GRU layers , use grid search to traverse these two hyperparameter combinations, preset learning rate The value range of is: ; in, is the learning rate The number of values; the preset number of GRU layers The value range of is: ; in, is the number of GRU layers The number of values ​​of ; S72. For each hyperparameter combination, The fold cross validation error formula calculates the error of this hyperparameter combination. fold cross validation error, The fold cross validation error formula is: ; in, For the hyperparameter combination Next, through The average validation error obtained by fold cross validation; For the The value of the learning rate; For the The value of the number of GRU layers; for The number of folds in the cross-validation; For a traversal of all The folded loop vector; For From 1 to The sum of all The validation errors of the folds are calculated to calculate their average value; For the Compromise using hyperparameter combinations Validation error after training the model; S73, through Fold restriction formula Number of folds in cross validation To restrict, The fold limit formula is: ; in, is the total number of hyperparameter combinations; is the number of hyperparameter types; is the floor rounding function; S74, select The hyperparameter combination with the smallest cross-validation error is taken as the optimal hyperparameter combination of the model.

[0029] For example, setting the value range of hyperparameters, the value range of learning rate is: 0.001, 0.01, 0.1; the value range of GRU layers is: 1, 2, 3; using the grid search method, traverse all possible hyperparameter combinations: Combination 1: learning rate = 0.001, GRU layer number = 1; Combination 2: learning rate = 0.001, number of GRU layers = 2; Combination 3: learning rate = 0.001, number of GRU layers = 3; Combination 4: learning rate = 0.01, number of GRU layers = 1; Combination 5: learning rate = 0.01, number of GRU layers = 2; Combination 6: learning rate = 0.01, number of GRU layers = 3; Combination 7: learning rate = 0.1, number of GRU layers = 1; Combination 8: learning rate = 0.1, number of GRU layers = 2; Combination 9: learning rate = 0.1, number of GRU layers = 3; Suppose you choose to perform 5-fold cross validation to evaluate the performance of each hyperparameter combination: For each hyperparameter combination, perform the following steps: Divide the dataset into 5 folds, for each fold, use 4 folds as training sets and the remaining 1 fold as validation set; Train the model and evaluate the error on the validation set; Repeat the above process 5 times, each time selecting a different fold as the validation set, and calculating the average of each validation error to obtain the fold cross validation error of the hyperparameter combination; assuming that when using combination 1, the validation errors obtained each time are: Verification error 1: 0.15; Verification error 2: 0.14; Verification error 3: 0.16; Verification error 4: 0.13; Verification error 5: 0.15; Then the fold cross validation error of combination 1 is: The cross validation error is 0.15+0.14+0.16+0.13+0.15 / 5=0.146. If the total number of hyperparameter combinations is 9 and the number of hyperparameter types is 2 (learning rate and number of GRU layers), use Discount limit formula to determine the reasonable discount number; The fold limit formula is: ; The fold limit here means that you can choose 5-fold cross validation to reduce the amount of calculation. Suppose after grid search and 5-fold cross validation, the following 5-fold cross validation error is obtained: Error of combination 1 = 0.146; Error of combination 2 = 0.143; Error of combination 3 = 0.147; Error of combination 4 = 0.141; Error of combination 5 = 0.139; Error of combination 6 = 0.144; Error of combination 7 = 0.148; Error of combination 8 = 0.146; Error of combination 9 = 0.142; The hyperparameter combination with the smallest error is selected as the optimal hyperparameter combination. In this example, combination 5 (learning rate = 0.01, number of GRU layers = 2) has the smallest error of 0.139, so it is the optimal hyperparameter combination finally selected.

[0030] The method of comparing the predicted audit risk coefficient of enterprise data assets with the preset audit risk coefficient threshold of enterprise data assets and evaluating the audit risk level faced by enterprise data assets includes: S81. Preset the audit risk coefficient threshold and limit the audit risk coefficient threshold through the threshold adjustment formula; the threshold adjustment formula is: ; in, is the audit risk factor threshold; is the median of historical audit risk factors; is the standard deviation of the historical audit risk coefficient; is the percentile of the historical audit risk factor; The adjustment factor for the difference between the median and percentile of the historical audit risk factor; S82. Adjustment coefficient for the difference between the median and percentile of historical audit risk coefficients through the coefficient adjustment model After adjustment, the coefficient adjustment model is: ; in, is the number of features that affect the historical audit risk factor; is the number of historical audit risk factors; is the variance of the historical audit risk coefficient; For example, the number of features that affect the historical audit risk factor is 8, the number of historical audit risk factors =2, and the variance of the historical audit risk coefficient is 1, then the adjustment coefficient of the difference between the median and percentile of the historical audit risk coefficient is About 0.5; S83, when When the value of is the first quartile, the audit risk coefficient threshold is recorded as ;when When the value of is the third quartile, the audit risk coefficient threshold is recorded as ; S84. According to the preset audit risk coefficient threshold, the audit risk level threshold faced by the enterprise data assets is further divided; the predicted audit risk coefficient is recorded as ; S85. Compare the predicted audit risk coefficient with the preset audit risk coefficient threshold to determine the audit risk level faced by the enterprise data assets; if , then the audit risk level faced by the enterprise data assets is judged to be low risk; if , then the audit risk level faced by the enterprise data assets is determined to be medium risk; if , then the audit risk level faced by the enterprise’s data assets is determined to be high risk.

[0031] For example, suppose the audit risk factor of a group of enterprise data assets is 0.7; according to the data set, the first quartile (Q1) is 0.4; the third quartile (Q3) is 0.8; therefore: T1 = 0.4 (first quartile); T2 = 0.8 (third quartile); Now assume that the predicted audit risk coefficient R is 0.7; according to the above rule: 0.4 < 0.7 <= 0.8, therefore, the audit risk level faced by enterprise data assets is medium risk.

[0032] According to the audit risk level faced by enterprise data assets, the methods for formulating corresponding asset optimization management strategies include: For enterprise data assets with low audit risk, the management strategies adopted include data quality inspection, data classification, hierarchical management, setting local data access control permissions and starting local data audit tracking mechanism; For enterprise data assets with moderate audit risk, the management strategies adopted include implementing joint data protection measures, strengthening data access control strategies, conducting comprehensive authority management in combination with role and attribute control, establishing data backup and recovery mechanisms, encrypting data storage and transmission, and implementing user behavior analysis and auditing, detecting abnormal data operations, and generating early warning information through enterprise intelligent management terminals; For enterprise data assets with high audit risks, we adopt a comprehensive optimization management strategy, conduct a comprehensive audit of all data, monitor all data operations, and generate audit reports through enterprise intelligent management terminals. We deploy DLP data leakage prevention technology, use high-strength encryption algorithms to encrypt data, and issue interception instructions through enterprise intelligent management terminals to automatically intercept unauthorized data transmission.

[0033] The preset enterprise data asset audit risk coefficient threshold is set by the staff, and the audit risk coefficients of different enterprise data assets are collected through the enterprise intelligent management terminal, and the average value of multiple enterprise data asset audit risk coefficients is taken as the preset enterprise data asset audit risk coefficient threshold; In this embodiment, the performance of the model on different data subsets can be effectively evaluated through cross-validation, ensuring the stability of the model training process; multi-fold validation can reduce the impact of the randomness of data division on the model evaluation results and improve the reliability of the results; the fold limit formula adjusts the fold according to the number of hyperparameter combinations to avoid unnecessary computational overhead; the model training is made more efficient, especially when there are many hyperparameter combinations, the computational cost can be significantly reduced; the standardized process of grid search and cross-validation methods makes the experimental results more repeatable; By introducing the threshold adjustment formula and combining the median, standard deviation, percentile and other statistical characteristics of the historical audit risk coefficient, the audit risk coefficient threshold can be adjusted dynamically; it can better adapt to the risk characteristics of different enterprises and industries, avoid the inflexibility caused by fixed thresholds, and make risk assessment more accurate and adaptable; using multiple statistical characteristics such as median, standard deviation, percentile, etc. to calculate the threshold of the audit risk coefficient makes the assessment method more scientific and comprehensive; it can reduce the impact of outliers on the results and provide more stable risk assessment results; The difference between the median and percentile of the historical audit risk coefficient is adjusted through the coefficient adjustment model, and the adjustment coefficient is further optimized using factors such as the number of features and variance. This can enhance the robustness and flexibility of the model, making the determination of risk thresholds more accurate. By using mathematical models and data analysis methods to determine the audit risk level, the subjectivity of the assessment process is reduced. This makes the risk assessment process more standardized, transparent and explainable, which helps to improve the objectivity and consistency of audit and management decisions.

[0034] Example 2 See also Figure 2 As shown, this embodiment is an enterprise data asset audit system based on big data, including: Data collection module, used to collect the company's financial information data, business operation data, management decision data and external industry data; The data processing module is used to pre-process the collected financial information data, business operation data, management decision data and external industry data to obtain financial information data set, business operation data set, management decision data set and external industry data set; Perform feature extraction and weighted fusion on financial information datasets, business operation datasets, management decision datasets, and external industry datasets to obtain a comprehensive feature dataset; The risk prediction module is used to build an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; The risk assessment module is used to compare the predicted audit risk coefficient of enterprise data assets with the preset enterprise data asset audit risk coefficient threshold to assess the audit risk level faced by enterprise data assets; The adjustment and optimization module is used to formulate corresponding optimized data management strategies based on the audit risk level faced by enterprise data assets.

[0035] Since the electronic device introduced in this embodiment is an electronic device used to implement a method and system for auditing enterprise data assets based on big data in the embodiment of this application, based on the method and system for auditing enterprise data assets based on big data introduced in the embodiment of this application, a person skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be described in detail here. As long as a person skilled in the art implements the electronic device used in a method and system for auditing enterprise data assets based on big data in the embodiment of this application, it belongs to the scope of protection of this application.

[0036] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0037] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for auditing enterprise data assets based on big data, characterized in that: include: S1. Collect the company's financial information data, business operation data, management decision data and external industry data; S2. Preprocess the collected financial information data, business operation data, management decision data and external industry data to obtain a financial information data set, a business operation data set, a management decision data set and an external industry data set; Perform feature extraction and weighted fusion on financial information datasets, business operation datasets, management decision datasets, and external industry datasets to obtain a comprehensive feature dataset; S3. Construct an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; S4. Compare the predicted audit risk coefficient of the enterprise data assets with the preset audit risk coefficient threshold of the enterprise data assets to evaluate the audit risk level faced by the enterprise data assets; S5. Develop corresponding asset optimization management strategies based on the audit risk level faced by enterprise data assets.

2. The enterprise data asset audit method based on big data according to claim 1 is characterized in that: The collected enterprise's financial information data includes: income and expenditure data, asset data, liability data, cash flow data and tax data; business operation data includes production data, inventory data, sales data, supply chain data, customer behavior data and operation performance data; Management decision data include strategic planning data, performance appraisal data, internal audit data, budget and forecast data and decision support data; external industry data include industry market data, economic environment data, policy and regulation data and technology trend data.

3. The enterprise data asset audit method based on big data according to claim 2 is characterized in that: The method of preprocessing the collected financial information data, business operation data, management decision data and external industry data to obtain the financial information data set, business operation data set, management decision data set and external industry data set includes: Use linear interpolation to fill missing values ​​in financial information data, business operation data, management decision data, and external industry data; use the linear relationship between the previous and next non-missing values ​​to calculate the estimated value of the missing value and replace the original missing value; check the uniqueness of financial information data, business operation data, management decision data, and external industry data, and remove duplicate records; Financial information data, business operation data, management decision data and external industry data are standardized and normalized, and data of different units and dimensions are converted into standard normal distribution to obtain financial information data set, business operation data set, management decision data set and external industry data set.

4. The enterprise data asset audit method based on big data according to claim 3 is characterized in that: The method of extracting features and weighted fusion of the financial information dataset, the business operation dataset, the management decision dataset and the external industry dataset to obtain a comprehensive feature dataset includes: The method for extracting features from a financial information data set is: calculating financial indicators and performing financial ratio analysis on the financial information data; performing time series analysis on the acquired financial indicators and extracting the trend, seasonality and cyclical features of the time series data; integrating the acquired financial indicators with the trend, seasonality and cyclical features of the data to obtain a financial feature data set; The method of feature extraction for business operation data set is: perform time series analysis on sales data in business operation data set to extract seasonal, trend and periodic features; perform production efficiency analysis on production data to extract productivity, equipment utilization and production cycle time features; calculate inventory turnover rate, inventory holding cost and supply and demand balance features for inventory data; perform customer feature analysis based on customer behavior data. The specific method is as follows: S41, extracting the RFM characteristics of the customer; calculating the R, F, and M indicators for each customer respectively, and scoring the R, F, and M indicators according to the calculation results; S42. Based on the R, F, and M indicator scores, the customer's scores in the three dimensions of R, F, and M are combined to obtain R, F, and M score combinations, each of which represents a different customer type, and a comprehensive score for each customer is calculated; The comprehensive score calculation formula is: ； in, For the Total score of customers; For the R-index score of each customer; For the F-index score of each customer; For the M-index score of each customer; For the R index score weight coefficient for each customer; For the The F-index score weight coefficient of each customer; For the The M indicator score weight coefficient for each customer; Index for customers; S43, preliminarily grouping the customers according to different score combinations, where each combination represents a different customer type, and restricting the score combinations through a score combination restriction model; The score combination restriction model is: ； in, is the proportion of a certain customer type in the total customer base; is the minimum comprehensive score among the customer groups; is the maximum comprehensive score among the customer groups; is the ratio of the scoring dimension to the total number of customers; is the total number of customers; Types of customers; is the number of scoring dimensions of RFM; S44, standardize the R, F, and M index scores of each customer; the standardization formula is: ； in, Score the standardized indicators; For customers R, F, M index scores; is the mean of the indicator scores; the standard deviation for scoring the indicator; Indexing for customers; The K-means clustering algorithm is used to further segment customers; the method of extracting features from the management decision data set is: SWOT analysis and key performance indicator analysis of the management decision data; The method of feature extraction for external industry data sets is: using time series analysis methods to analyze market trends and extract macroeconomic features; Through the above steps, the financial feature data set, business operation feature data set, management decision feature data set and external industry feature data set after feature extraction are obtained; the financial feature data set, business operation feature data set, management decision feature data set and external industry feature data set after feature extraction are weighted fused to obtain a comprehensive feature data set.

5. The enterprise data asset audit method based on big data according to claim 4 is characterized in that: The method of performing weighted fusion on the feature-extracted financial feature data set, business operation feature data set, management decision feature data set and external industry feature data set to obtain a comprehensive feature data set includes: The financial feature dataset, business operation feature dataset, management decision feature dataset and external industry feature dataset after feature extraction are fused through a weighted model to form a comprehensive feature dataset. The financial feature dataset is recorded as , the business operation feature dataset is recorded as , the management decision feature dataset is recorded as ; The external industry characteristic data set is recorded as ; The weighted model is: ； in, is the weight coefficient of the financial feature data set; is the weight coefficient of the business operation feature data set; is the weight coefficient of the management decision feature data set; is the weight coefficient of the external industry characteristic data set.

6. The enterprise data asset audit method based on big data according to claim 5 is characterized in that: The training method of the audit risk prediction model includes: The data set is divided into a training set, a validation set and a test set. The data set includes a historical comprehensive feature data set and the audit risk coefficient of the corresponding enterprise asset data; Construct an audit risk prediction model, which includes an input layer, a GRU layer, a fully connected layer, and an output layer; the input layer of the audit risk prediction model is used to input a historical comprehensive feature data set, and the number of neurons in the input layer matches the number of features in the historical comprehensive feature data set; The GRU layer is used to process the historical comprehensive feature data set, and the number of GRU layers and neurons are adjusted according to the complexity of the task; the fully connected layer is used to provide additional nonlinear transformation; the model output layer is used to output the network signal performance index, and the predicted value is output through a neuron, and the identity function is used as the activation function; the audit risk prediction model is a gated recurrent unit model; The mean absolute error is used as the loss function to measure the error between the model's predicted value and the actual value; the mean absolute error loss function is: ； in, is the mean absolute error; is the number of data sets; For the The true value of the audit risk coefficient of the enterprise asset data; For the The predicted value of the audit risk coefficient of the asset data of each enterprise; Use the training set data to train the model and minimize the loss function using the Adam optimizer. Use the validation set to evaluate the performance of the model and tune the model's hyperparameters until the model performance no longer improves significantly or reaches the preset stopping condition. Use the test set to evaluate the performance of the model in the prediction task, input the current comprehensive feature data set into the trained audit risk prediction model, and obtain the audit risk coefficient of the enterprise data assets.

7. The enterprise data asset audit method based on big data according to claim 6 is characterized in that: The method for tuning the hyperparameters of the model includes: S71, there are two preset hyperparameters: learning rate and the number of GRU layers , use grid search to traverse these two hyperparameter combinations, preset learning rate The value range of is: ;in, is the learning rate The number of values; the preset number of GRU layers The value range of is: ； in, is the number of GRU layers The number of values ​​of ; S72. For each hyperparameter combination, The fold cross validation error formula calculates the error of this hyperparameter combination. fold cross validation error, The fold cross validation error formula is: ； in, For the hyperparameter combination Next, through The average validation error obtained by fold cross validation; For the The value of the learning rate; For the The value of the number of GRU layers; for The number of folds in the cross-validation; For a traversal of all The folded loop vector; For From 1 to The sum of For the Compromise using hyperparameter combinations Validation error after training the model; S73, through Fold restriction formula Number of folds in cross validation To restrict The fold limit formula is: ； in, is the total number of hyperparameter combinations; is the number of hyperparameter types; is the floor rounding function; S74, select The hyperparameter combination with the smallest cross-validation error is taken as the optimal hyperparameter combination of the model.

8. The enterprise data asset audit method based on big data according to claim 7 is characterized in that: The method of comparing the predicted audit risk coefficient of the enterprise data asset with the preset enterprise data asset audit risk coefficient threshold to evaluate the audit risk level faced by the enterprise data asset includes: S81. Preset the audit risk coefficient threshold and limit the audit risk coefficient threshold through the threshold adjustment formula; the threshold adjustment formula is: ； in, is the audit risk factor threshold; is the median of historical audit risk factors; is the standard deviation of the historical audit risk coefficient; is the percentile of the historical audit risk factor; The adjustment factor for the difference between the median and percentile of the historical audit risk factor; S82. Adjustment coefficient for the difference between the median and percentile of historical audit risk coefficients through the coefficient adjustment model Adjustment is performed, and the coefficient adjustment model is: ； in, is the number of features that affect the historical audit risk factor; is the number of historical audit risk factors; is the variance of the historical audit risk coefficient; S83, when When the value of is the first quartile, the audit risk coefficient threshold is recorded as ;when When the value of is the third quartile, the audit risk coefficient threshold is recorded as ; S84. According to the preset audit risk coefficient threshold, the audit risk level threshold faced by the enterprise data assets is further divided; the predicted audit risk coefficient is recorded as ; S85. Compare the predicted audit risk coefficient with the preset audit risk coefficient threshold to determine the audit risk level faced by the enterprise data assets; if , then the audit risk level faced by the enterprise data assets is judged to be low risk; if , then the audit risk level faced by the enterprise data assets is determined to be medium risk; if , then the audit risk level faced by the enterprise’s data assets is determined to be high risk.

9. The enterprise data asset audit method based on big data according to claim 8 is characterized in that: The method for formulating a corresponding asset optimization management strategy according to the audit risk level faced by the enterprise data assets includes: For enterprise data assets with low audit risk, the management strategies adopted include data quality inspection, data classification, hierarchical management, setting local data access control permissions and starting local data audit tracking mechanism; For enterprise data assets with moderate audit risk, the management strategies adopted include implementing joint data protection measures, strengthening data access control strategies, conducting comprehensive authority management in combination with role and attribute control, establishing data backup and recovery mechanisms, encrypting data storage and transmission, and implementing user behavior analysis and auditing, detecting abnormal data operations, and generating early warning information through enterprise intelligent management terminals; For enterprise data assets with high audit risks, we adopt a comprehensive optimization management strategy, conduct a comprehensive audit of all data, monitor all data operations, and generate audit reports through enterprise intelligent management terminals. We deploy DLP data leakage prevention technology, use high-strength encryption algorithms to encrypt data, and issue interception instructions through enterprise intelligent management terminals to automatically intercept unauthorized data transmission.

10. An enterprise data asset audit system based on big data, used to implement the enterprise data asset audit method based on big data as claimed in any one of claims 1 to 9, characterized in that: include: Data collection module, used to collect the company's financial information data, business operation data, management decision data and external industry data; The data processing module is used to pre-process the collected financial information data, business operation data, management decision data and external industry data to obtain financial information data set, business operation data set, management decision data set and external industry data set; Perform feature extraction and weighted fusion on financial information datasets, business operation datasets, management decision datasets, and external industry datasets to obtain a comprehensive feature dataset; The risk prediction module is used to build an audit risk prediction model, input the comprehensive feature data set into the audit risk prediction model, and predict the audit risk coefficient of the enterprise data assets; The risk assessment module is used to compare the predicted audit risk coefficient of enterprise data assets with the preset enterprise data asset audit risk coefficient threshold to assess the audit risk level faced by enterprise data assets; The adjustment and optimization module is used to formulate corresponding optimized data management strategies according to the audit risk level faced by the enterprise data assets; each module is connected by wired and / or wireless means.

Images