Breakdown data analysis method and system based on big data

Abstract

The invention provides a big data-based production breaking data analysis method and system, and the method comprises the steps: S1, obtaining company financial data, production data and market emotion data, and combining the three pieces of data as first data; s2, preprocessing the first data to obtain second data, and storing the second data in a database; s3, dividing the market emotion data in the second data into positive emotions and negative emotions, calculating emotion evaluation coefficients according to the number of days to obtain an emotion evaluation coefficient trend table, and inserting financial data and production data into the table to obtain a company operation situation trend table; and S4, carrying out production breaking analysis on the company operation condition trend table, carrying out quantitative evaluation on the enterprise production breaking risk according to an analysis result, and when the evaluation value of the production breaking risk exceeds a threshold value, giving an alarm by an early warning device. According to the invention, through big data analysis, potential production breaking risk enterprises can be identified more accurately, and early warning is provided for creditor, court and production breaking managers.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to a bankruptcy data analysis method and system based on big data. Background Art

[0002] In the modern business environment, data has become an important asset for enterprises. Enterprises need to use data analysis to optimize business processes, improve customer experience, discover market opportunities, and reduce operating costs. As an important application field of data analysis, the demand for bankruptcy data analysis is also increasing with the intensification of business competition. Enterprises and organizations need to use big data analysis technology to better understand and respond to market changes, reduce bankruptcy risks, and improve operating efficiency. Summary of the invention

[0003] Based on the increasing demand, the present invention proposes a bankruptcy data analysis method and system based on big data, and the technical solutions adopted are as follows:

[0004] A bankruptcy data analysis method based on big data, characterized in that the method comprises:

[0005] S1: Obtain company financial data, production data, and market sentiment data, and combine the three data as the first data;

[0006] S2: preprocessing the first data to obtain second data, and storing the second data in a database;

[0007] S3: Divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain a sentiment evaluation coefficient trend table, insert financial data and production data into the table, and obtain a company operation situation trend table;

[0008] S4: Conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessed value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

[0009] Preferably, the S1 includes:

[0010] S11: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss number in the distributed ledger through the API interface;

[0011] S12: Obtain the production log of the production equipment, and collect the production data of the production equipment in real time through the production log;

[0012] S13: Use crawler systems to collect relevant text data on social platforms as a data source for market sentiment analysis;

[0013] S14: Merge the financial data acquired in S11, the production data acquired in S12, and the market sentiment data acquired in S13, and use the merged data as the first data.

[0014] Preferably, the S2 includes:

[0015] S21: After cleaning the first data, standardize the first data;

[0016] S22: Construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: a generator learns the distribution of real data and generates simulated data similar to the real data. A judge then judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in a database.

[0017] Preferably, S3 includes:

[0018] S31: setting emotion indicators, which include positive emotion indicators and negative emotion indicators, calculating emotion evaluation coefficients on a daily basis, and drawing an emotion evaluation coefficient trend chart based on the emotion evaluation coefficients;

[0019] S32: Insert the financial data and operation data in the second data into the emotion evaluation coefficient trend table, and align the time to ensure that the emotion evaluation coefficient and the financial data and production data are aligned at the same time to obtain the company's operation status trend table.

[0020] Preferably, the S4 includes:

[0021] S41: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results;

[0022] S42: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold value. If the assessment value corresponding to the risk assessment result is higher than the threshold value, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

[0023] A bankruptcy data analysis system based on big data, characterized in that the system comprises:

[0024] Data collection system: obtain company financial data, production data and market sentiment data, and combine the three data as the first data;

[0025] Data preprocessing system: preprocessing the first data to obtain second data, and storing the second data in a database;

[0026] Data analysis system: divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain the sentiment evaluation coefficient trend table, insert the financial data and production data into the table, and obtain the company's operating situation trend table;

[0027] Risk assessment system: conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessment value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

[0028] Preferably, the data acquisition system comprises:

[0029] Financial information extraction system: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss in the distributed ledger through the API interface;

[0030] Production information extraction system: obtain the production log of production equipment and collect the production data of production equipment in real time through the production log;

[0031] Crawler system: Use the crawler system to collect relevant text data on social platforms as a data source for market sentiment analysis;

[0032] Data merging system: merges the financial data acquired by S11, the production data acquired by S12, and the market sentiment data acquired by S13, and uses the merged data as the first data.

[0033] Preferably, the data preprocessing system comprises:

[0034] Data standardization system: after cleaning the first data, standardize the first data;

[0035] Simulated data training system: construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: the generator learns the distribution of real data and generates simulated data similar to the real data. The judge then judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in the database.

[0036] Preferably, the data analysis system comprises:

[0037] Emotional scoring system: setting emotional indicators, including positive emotional indicators and negative emotional indicators, calculating emotional evaluation coefficients on a daily basis, and drawing an emotional evaluation coefficient trend chart based on the emotional evaluation coefficients;

[0038] Time synchronization system: insert the financial data and operation data in the second data into the sentiment evaluation coefficient trend table, and align the time to ensure that the sentiment evaluation coefficient, financial data and production data are aligned at the same time to obtain the company's operation status trend table.

[0039] Preferably, the risk assessment system comprises:

[0040] Risk assessment model: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results;

[0041] Early warning system: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold. If the assessment value corresponding to the risk assessment result is higher than the threshold, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

[0042] Beneficial effects of the present invention: Through big data analysis, potential bankruptcy risk enterprises can be more accurately identified, providing early warnings for creditors, courts and bankruptcy administrators. Using data analysis tools, a comprehensive assessment of the financial status and operating capacity of bankrupt enterprises can be conducted to provide a scientific basis for decision-making. According to the results of bankruptcy data analysis, the allocation of claims, asset disposal and other plans can be optimized to maximize the use of resources. Timely discover and prevent legal risks, financial risks, etc. that may arise during the bankruptcy process to protect the interests of all parties. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a bankruptcy data analysis method based on big data described in the present invention. DETAILED DESCRIPTION

[0044] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0045] One embodiment of the present invention is a bankruptcy data analysis method based on big data, characterized in that the method comprises:

[0046] S1: Obtain company financial data, production data, and market sentiment data, and combine the three data as the first data;

[0047] S2: preprocessing the first data to obtain second data, and storing the second data in a database;

[0048] S3: Divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain a sentiment evaluation coefficient trend table, insert financial data and production data into the table, and obtain a company operation situation trend table;

[0049] S4: Conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessed value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

[0050] The working principle and effect of the above technical solution are: financial data is obtained through multiple channels such as the company's internal financial system, operation management system, and market financial data collection platform. These data include but are not limited to financial statements, operating indicators, market trends, industry reports, etc., as the first data. The first data is cleaned, converted and integrated to remove duplicate, erroneous and invalid data, unify the data format and standard, and ensure the accuracy and consistency of the data. The pre-processed second data is stored in a distributed file system. The second data is deeply analyzed, and key information is extracted using technical means such as data mining, statistical analysis, and machine learning to form third data. The third data is analyzed in multiple dimensions and angles, including but not limited to financial health analysis, operational efficiency analysis, market competitiveness analysis, risk factor identification, etc., to reveal the company's potential bankruptcy risk. According to the analysis results, a bankruptcy risk quantification model is established to quantify the bankruptcy risk of the enterprise. Quantitative scoring should take into account multiple factors, such as the degree of deterioration of financial indicators, the speed of decline in operational efficiency, and the weakening of market competitiveness. When the bankruptcy risk quantitative evaluation coefficient exceeds the preset threshold, the early warning device automatically issues a warning. Warning information should clearly indicate the risk points, risk levels and possible consequences so that enterprises can take timely measures to deal with the risks. Collecting data from multiple channels, including internal financial and operational data of the company and market financial data, can ensure the comprehensiveness and accuracy of the data and provide a solid foundation for bankruptcy risk assessment. Using intelligent analysis technologies such as data mining, statistical analysis, and machine learning, data processing and analysis can be automated to improve the accuracy and efficiency of analysis. Compared with traditional manual analysis methods, this method can identify potential risk points more quickly. By quantitatively scoring, the bankruptcy risk is converted into specific values, which makes it easier for enterprises to intuitively understand their own risk status. At the same time, intuitive charts and reports can be generated to help decision makers better understand the risk distribution and changing trends. When the bankruptcy risk exceeds the preset threshold, the early warning device can quickly issue a warning to remind the enterprise to take timely measures to deal with the risk. This helps enterprises to formulate risk response plans in advance and reduce losses caused by bankruptcy risks.

[0051] In one embodiment of the present invention, the S1 includes:

[0052] S11: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss number in the distributed ledger through the API interface;

[0053] S12: Obtain the production log of the production equipment, and collect the production data of the production equipment in real time through the production log;

[0054] S13: Use crawler systems to collect relevant text data on social platforms as a data source for market sentiment analysis;

[0055] S14: Merge the financial data acquired in S11, the production data acquired in S12, and the market sentiment data acquired in S13, and use the merged data as the first data.

[0056] The working principle and effect of the above technical solution are as follows: set up special nodes on the blockchain network, which are responsible for extracting data from the blockchain distributed ledger. The distributed ledger technology of the blockchain ensures the immutability and transparency of the data, making the acquired data highly credible. By calling the API interface provided by the blockchain platform, the financial data stored on the blockchain can be easily obtained. Write a special crawler program to regularly visit the social platform and capture relevant text data, clean and process the captured text data, remove useless information and noise data, and extract useful information points. These information points can be used to analyze the market's views on the company, the reputation of the product, and potential market risks. Integrate the data collected in steps S11, S12, and S13. These data cover multiple aspects such as the company's financial status, production operations, and market feedback, providing comprehensive data support for bankruptcy risk analysis. Sensors can collect status information of production equipment in real time, provide real-time monitoring capabilities for the production process, and help to discover and solve problems in a timely manner.

[0057] In one embodiment of the present invention, S2 includes:

[0058] S21: After cleaning the first data, standardize the first data;

[0059] S22: Construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: a generator learns the distribution of real data and generates simulated data similar to the real data. A judge then judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in a database.

[0060] The working principle and effect of the above technical solution are as follows: Data cleaning is a key step in data preprocessing, which aims to improve data quality and lay a solid foundation for subsequent modeling and analysis. For the first data, data cleaning may include operations such as missing value processing, outlier detection and processing, noise data smoothing, and data duplication processing. By deleting, filling or correcting these problems in the data, the accuracy and completeness of the data can be ensured. Standardization processing is to convert the data feature values ​​to the same order of magnitude or range to eliminate the influence of different dimensions and orders of magnitude on data analysis. Generative adversarial network (GAN) is a deep learning model consisting of two sub-networks, generator and discriminator. The goal of the generator is to learn the distribution of real data and generate simulated data as close to the real data as possible; while the goal of the discriminator is to distinguish whether the input data is real or simulated data generated by the generator. The two confront each other during the training process and continuously optimize their own performance until the generator can generate simulated data that is indistinguishable from the real thing. Initialization: The generator and discriminator are initialized to random states, usually with random weights.

[0061] Generator training: In each iteration, the generator receives random noise as input and tries to generate simulated data. The discriminator evaluates these simulated data and gives the probability that it is real data. The generator adjusts its parameters based on the feedback from the discriminator to generate more realistic data.

[0062] Discriminator training: At the same time, the discriminator also receives real data from the real dataset and simulated data generated by the generator as input. Its goal is to distinguish whether these data are real or simulated, and optimize its parameters to improve classification accuracy.

[0063] Alternating training: The training process of the generator and the discriminator is carried out alternately until a predetermined number of training rounds is reached or other stopping conditions are met. In this process, the generator gradually learns to generate more realistic simulated data, while the discriminator finds it increasingly difficult to distinguish between real data and simulated data.

[0064] After the training is completed, the simulated data generated by the generator can be added to the first data to form the second data. These simulated data can be used as a supplement to the real data to enhance the diversity and scale of the data set. Finally, the second data is stored in the database for subsequent analysis and modeling. The simulated data generated by the adversarial network model can be used as a supplement to the real data, thereby expanding the scale of the data set. This method is particularly effective when the amount of data is insufficient or the data distribution is uneven. Increasing the scale of the data set can improve the generalization ability and stability of the model. The addition of simulated data can increase the diversity of the data set, allowing the model to learn more different data patterns and features. This helps the model better adapt to complex and changing data environments and improves the prediction accuracy and robustness of the model. The development and application of new technologies such as generative adversarial networks provide new ideas and methods for data analysis and modeling. These new technologies can handle more complex and diverse data and promote innovation and development in the fields of data science and artificial intelligence.

[0065] In one embodiment of the present invention, S3 includes:

[0066] S31: setting emotion indicators, which include positive emotion indicators and negative emotion indicators, calculating emotion evaluation coefficients on a daily basis, and drawing an emotion evaluation coefficient trend chart based on the emotion evaluation coefficients;

[0067] S32: Insert the financial data and operation data in the second data into the emotion evaluation coefficient trend table, and align the time to ensure that the emotion evaluation coefficient and the financial data and production data are aligned at the same time to obtain the company's operation status trend table.

[0068] The working principle and effect of the above technical solution are as follows: Use natural language processing tools to perform sentiment analysis on each text data, calculate its positive and negative sentiment evaluation coefficients, and get +1 point if it is positive sentiment, and -1 point if it is negative sentiment. Summarize the positive sentiment evaluation coefficients and negative sentiment evaluation coefficients of each day, and calculate the net sentiment evaluation coefficient of each day. Combine the financial data and operational data in the second data with the sentiment index, and align the time to ensure that the sentiment index and financial operational data are aligned at the same time. The second integration system is responsible for integrating the sentiment index data output by the sentiment scoring system with the financial data and operational data processed by the time synchronization system. Conduct in-depth analysis of the third data to generate key indicators and visualizations for reporting and further decision-making. Through the real-time processing of social media and customer feedback comments by sentiment analysis tools, valuable sentiment information can be obtained in a timely manner, and the sentiment changes in the market can be quickly understood. The combination of sentiment, financial and operational data provides a panoramic view, covering information from multiple dimensions inside and outside the enterprise, which helps to more comprehensively understand the overall performance of the enterprise. Through comprehensive analysis of the three dimensions of emotion, finance and operation, corporate risks can be predicted and managed more accurately. In particular, emotional indicators can provide early warning of potential financial or operational crises.

[0069] In one embodiment of the present invention, the S4 includes:

[0070] S41: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results;

[0071] S42: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold value. If the assessment value corresponding to the risk assessment result is higher than the threshold value, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

[0072] The working principle and effect of the above technical solution are as follows: The risk assessment model first needs to identify and investigate potential risk factors. After identifying the risk factors, the risk assessment model will conduct a detailed analysis and evaluation of these factors. This includes determining the possibility of risk occurrence and the severity it may bring, as well as evaluating the impact of these risks on corporate goals. In order to more intuitively compare and sort the importance and priority of different risks, the risk assessment model will quantify the risk factors. Taking into account the probability and impact of risks, the risk assessment model will evaluate the risks. The early warning system will set a series of risk thresholds based on the actual situation of the enterprise and risk management needs. These thresholds represent the upper limit of the risk level that the enterprise can bear. The early warning system will monitor the operating conditions of the enterprise and changes in the external environment in real time, and collect data and information related to bankruptcy risks. The data collected in real time is input into the risk assessment model to obtain the current risk assessment results. The early warning system will then compare these results with the preset thresholds. If the risk assessment results exceed the preset threshold, the early warning system will immediately issue an alarm. This is done through email, text messages, internal system notifications, etc. to ensure that the enterprise can receive risk warning information in a timely manner. Once the risk warning information is received, the enterprise needs to quickly initiate an emergency response mechanism. This includes organizing relevant personnel to hold emergency meetings, developing response strategies, and allocating resources to minimize the impact of bankruptcy risks on the enterprise. The risk assessment model can systematically identify, analyze and quantify the various risks faced by the enterprise, including economic, legal, technological, market, supply chain and other aspects in the internal and external environment. This comprehensive risk assessment helps enterprises to understand their own situation more comprehensively and provide strong support for the formulation of strategies and plans. By using mathematical models and statistical methods, the risk assessment model can quantify risk factors and make the risk assessment results more objective and scientific. This scientific assessment helps enterprises to more accurately judge the importance and priority of risks, so as to formulate risk management strategies in a targeted manner. The risk assessment model can be dynamically adjusted as the operating conditions and external environment of the enterprise change. When new risk factors appear, the model can be incorporated and analyzed and evaluated in a timely manner to ensure that the enterprise always maintains its sensitivity to risks and its ability to respond. The early warning system has a high degree of automation and can automatically operate and issue alarms without human intervention. This automated feature reduces the burden of manual monitoring and improves the efficiency and accuracy of early warning. The early warning system can be customized and adjusted according to the actual needs of the enterprise. Enterprises can set different thresholds and alarm methods based on their own risk tolerance and management needs to meet personalized risk management needs.

[0073] One embodiment of the present invention is a bankruptcy data analysis system based on big data, characterized in that the system includes:

[0074] Data collection system: obtain company financial data, production data and market sentiment data, and combine the three data as the first data;

[0075] Data preprocessing system: preprocessing the first data to obtain second data, and storing the second data in a database;

[0076] Data analysis system: divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain the sentiment evaluation coefficient trend table, insert the financial data and production data into the table, and obtain the company's operating situation trend table;

[0077] Risk assessment system: conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessment value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

[0078] The working principle and effect of the above technical solution are: financial data is obtained through multiple channels such as the company's internal financial system, operation management system, and market financial data collection platform. These data include but are not limited to financial statements, operating indicators, market trends, industry reports, etc., as the first data. The first data is cleaned, converted and integrated to remove duplicate, erroneous and invalid data, unify the data format and standard, and ensure the accuracy and consistency of the data. The pre-processed second data is stored in a distributed file system. The second data is deeply analyzed, and key information is extracted using technical means such as data mining, statistical analysis, and machine learning to form third data. The third data is analyzed in multiple dimensions and angles, including but not limited to financial health analysis, operational efficiency analysis, market competitiveness analysis, risk factor identification, etc., to reveal the company's potential bankruptcy risk. According to the analysis results, a bankruptcy risk quantification model is established to quantify the bankruptcy risk of the enterprise. Quantitative scoring should take into account multiple factors, such as the degree of deterioration of financial indicators, the speed of decline in operational efficiency, and the weakening of market competitiveness. When the bankruptcy risk quantitative evaluation coefficient exceeds the preset threshold, the early warning device automatically issues a warning. Warning information should clearly indicate the risk points, risk levels and possible consequences so that enterprises can take timely measures to deal with the risks. Collecting data from multiple channels, including internal financial and operational data of the company and market financial data, can ensure the comprehensiveness and accuracy of the data and provide a solid foundation for bankruptcy risk assessment. Using intelligent analysis technologies such as data mining, statistical analysis, and machine learning, data processing and analysis can be automated to improve the accuracy and efficiency of analysis. Compared with traditional manual analysis methods, this method can identify potential risk points more quickly. By quantitatively scoring, the bankruptcy risk is converted into specific values, which makes it easier for enterprises to intuitively understand their own risk status. At the same time, intuitive charts and reports can be generated to help decision makers better understand the risk distribution and changing trends. When the bankruptcy risk exceeds the preset threshold, the early warning device can quickly issue a warning to remind the enterprise to take timely measures to deal with the risk. This helps enterprises to formulate risk response plans in advance and reduce losses caused by bankruptcy risks.

[0079] In one embodiment of the present invention, the data acquisition system comprises:

[0080] Financial information extraction system: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss in the distributed ledger through the API interface;

[0081] Production information extraction system: obtain the production log of production equipment and collect the production data of production equipment in real time through the production log;

[0082] Crawler system: Use the crawler system to collect relevant text data on social platforms as a data source for market sentiment analysis;

[0083] Data merging system: merges the financial data acquired by S11, the production data acquired by S12, and the market sentiment data acquired by S13, and uses the merged data as the first data.

[0084] The working principle and effect of the above technical solution are as follows: set up special nodes on the blockchain network, which are responsible for extracting data from the blockchain distributed ledger. The distributed ledger technology of the blockchain ensures the immutability and transparency of the data, making the acquired data highly credible. By calling the API interface provided by the blockchain platform, the financial data stored on the blockchain can be easily obtained. Write a special crawler program to regularly visit the social platform and capture relevant text data, clean and process the captured text data, remove useless information and noise data, and extract useful information points. These information points can be used to analyze the market's views on the company, the reputation of the product, and potential market risks. Integrate the data collected in steps S11, S12, and S13. These data cover multiple aspects such as the company's financial status, production operations, and market feedback, providing comprehensive data support for bankruptcy risk analysis. Sensors can collect status information of production equipment in real time, provide real-time monitoring capabilities for the production process, and help to discover and solve problems in a timely manner.

[0085] In one embodiment of the present invention, the data preprocessing system comprises:

[0086] Data standardization system: after cleaning the first data, standardize the first data;

[0087] Simulated data training system: construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: the generator learns the distribution of real data and generates simulated data similar to the real data, and then the judge judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in the database. The working principle and effect of the above technical solution are: data cleaning is a key step in data preprocessing, which aims to improve data quality and lay a solid foundation for subsequent modeling and analysis. For the first data, data cleaning may include operations such as missing value processing, outlier detection and processing, noise data smoothing, and data duplication processing. By deleting, filling or correcting these problems in the data, the accuracy and completeness of the data can be ensured. Standardization processing is to convert the data feature values ​​to the same order of magnitude or range to eliminate the influence of different dimensions and orders of magnitude on data analysis. Generative adversarial network (GAN) is a deep learning model consisting of two sub-networks: generator and discriminator. The goal of the generator is to learn the distribution of real data and generate simulated data that is as close to the real data as possible; while the goal of the discriminator is to distinguish whether the input data is real or simulated data generated by the generator. The two compete with each other during the training process and continuously optimize their own performance until the generator can generate simulated data that is indistinguishable from the real thing. Initialization: The generator and discriminator are initialized to random states, usually with random weights.

[0088] Generator training: In each iteration, the generator receives random noise as input and tries to generate simulated data. The discriminator evaluates these simulated data and gives the probability that it is real data. The generator adjusts its parameters based on the feedback from the discriminator to generate more realistic data.

[0089] Discriminator training: At the same time, the discriminator also receives real data from the real dataset and simulated data generated by the generator as input. Its goal is to distinguish whether these data are real or simulated, and optimize its parameters to improve classification accuracy.

[0090] Alternating training: The training process of the generator and the discriminator is carried out alternately until a predetermined number of training rounds is reached or other stopping conditions are met. In this process, the generator gradually learns to generate more realistic simulated data, while the discriminator finds it increasingly difficult to distinguish between real data and simulated data.

[0091] After the training is completed, the simulated data generated by the generator can be added to the first data to form the second data. These simulated data can be used as a supplement to the real data to enhance the diversity and scale of the data set. Finally, the second data is stored in the database for subsequent analysis and modeling. The simulated data generated by the adversarial network model can be used as a supplement to the real data, thereby expanding the scale of the data set. This method is particularly effective when the amount of data is insufficient or the data distribution is uneven. Increasing the scale of the data set can improve the generalization ability and stability of the model. The addition of simulated data can increase the diversity of the data set, allowing the model to learn more different data patterns and features. This helps the model better adapt to complex and changing data environments and improves the prediction accuracy and robustness of the model. The development and application of new technologies such as generative adversarial networks provide new ideas and methods for data analysis and modeling. These new technologies can handle more complex and diverse data and promote innovation and development in the fields of data science and artificial intelligence.

[0092] In one embodiment of the present invention, the data analysis system comprises:

[0093] Emotional scoring system: setting emotional indicators, including positive emotional indicators and negative emotional indicators, calculating emotional evaluation coefficients on a daily basis, and drawing an emotional evaluation coefficient trend chart based on the emotional evaluation coefficients;

[0094] Time synchronization system: insert the financial data and operation data in the second data into the sentiment evaluation coefficient trend table, and align the time to ensure that the sentiment evaluation coefficient, financial data and production data are aligned at the same time to obtain the company's operation status trend table.

[0095] The working principle and effect of the above technical solution are as follows: Use natural language processing tools to perform sentiment analysis on each text data, calculate its positive and negative sentiment evaluation coefficients, and get +1 point if it is positive sentiment, and -1 point if it is negative sentiment. Summarize the positive sentiment evaluation coefficients and negative sentiment evaluation coefficients of each day, and calculate the net sentiment evaluation coefficient of each day. Combine the financial data and operational data in the second data with the sentiment index, and align the time to ensure that the sentiment index and financial operational data are aligned at the same time. The second integration system is responsible for integrating the sentiment index data output by the sentiment scoring system with the financial data and operational data processed by the time synchronization system. Conduct in-depth analysis of the third data to generate key indicators and visualizations for reporting and further decision-making. Through the real-time processing of social media and customer feedback comments by sentiment analysis tools, valuable sentiment information can be obtained in a timely manner, and the sentiment changes in the market can be quickly understood. The combination of sentiment, financial and operational data provides a panoramic view, covering information from multiple dimensions inside and outside the enterprise, which helps to more comprehensively understand the overall performance of the enterprise. Through comprehensive analysis of the three dimensions of emotion, finance and operation, corporate risks can be predicted and managed more accurately. In particular, emotional indicators can provide early warning of potential financial or operational crises.

[0096] In one embodiment of the present invention, the risk assessment system comprises:

[0097] Risk assessment model: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results;

[0098] Early warning system: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold. If the assessment value corresponding to the risk assessment result is higher than the threshold, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

[0099] The risk assessment result is obtained by the following formula:

[0100]

[0101] Among them, R(t) represents the risk assessment result on day t, ΔE t represents the change in the emotion evaluation coefficient on the tth day, E t-1 represents the sentiment evaluation coefficient on day t-1, represents the year-on-year change in profit on day t, represents the month-on-month change in profit on the tth day, S(F) represents the standard deviation of profit change within a month, and O t represents the production efficiency index on day t;

[0102] The year-on-year change in the profit on the tth day, the month-on-month change in the profit on the tth day, and the standard deviation of the profit change within one month are obtained by the following formula:

[0103]

[0104]

[0105]

[0106] Among them, F t represents the profit of the enterprise on day t, F t ' represents the profit of the enterprise on the same day last year, F t " represents the profit of the enterprise on the same day last month, and μ represents the average profit of the enterprise this month.

[0107] Furthermore, the O t Obtained through the following calculation formula:

[0108]

[0109] Where T represents the usage time of the production equipment within the t-th day, and N represents the number of qualified products produced by the production equipment within the t-th day;

[0110] The larger the value of R(t), the greater the risk of bankruptcy faced by the enterprise. When R(t) exceeds the set threshold, the early warning device will sound an alarm to warn the enterprise that there is a risk of bankruptcy.

[0111] The working principle and effect of the above technical solution are as follows: The risk assessment model first needs to identify and investigate potential risk factors. After identifying the risk factors, the risk assessment model will conduct a detailed analysis and evaluation of these factors. This includes determining the likelihood of the risk occurring and the severity it may bring, as well as evaluating the impact of these risks on corporate goals. In order to more intuitively compare and sort the importance and priority of different risks, the risk assessment model will quantify the risk factors. Taking into account the probability and impact of the risk, the risk assessment model will evaluate the risk. In the formula, Can capture changes in market sentiment in a timely manner. It can quickly reflect short-term profit changes and is very effective in capturing sudden risks. The use of logarithmic function calculations can reduce the interference of volatility and provide more accurate short-term risk assessment. It usually directly affects the comprehensive competitiveness and profitability of the enterprise and is a direct reflection of the operating status. The early warning system will set a series of risk thresholds based on the actual situation of the enterprise and the risk management needs. These thresholds represent the upper limit of the risk level that the enterprise can bear. The early warning system will monitor the operating status of the enterprise and the changes in the external environment in real time, and collect data and information related to bankruptcy risk. The data collected in real time is input into the risk assessment model to obtain the current risk assessment results. The early warning system will then compare these results with the preset thresholds. If the risk assessment results exceed the preset thresholds, the early warning system will immediately issue an alarm. This is done through emails, text messages, internal system notifications, etc. to ensure that the enterprise can receive risk warning information in a timely manner. Once the risk warning information is received, the enterprise needs to quickly initiate an emergency response mechanism. This includes organizing relevant personnel to hold emergency meetings, formulating response strategies, allocating resources, etc., to minimize the impact of bankruptcy risks on the enterprise. The risk assessment model can systematically identify, analyze and quantify the various risks faced by the enterprise, including economic, legal, technical, market, supply chain and other aspects in the internal and external environment. This comprehensive risk assessment helps enterprises to have a more comprehensive understanding of their own situation and provide strong support for the formulation of strategies and plans. By using mathematical models and statistical methods, risk assessment models can quantify risk factors, making risk assessment results more objective and scientific. This scientific assessment helps enterprises to more accurately judge the importance and priority of risks, so as to formulate risk management strategies in a targeted manner. Risk assessment models can be dynamically adjusted as the operating conditions and external environment of the enterprise change. When new risk factors emerge, the model can be incorporated and analyzed and evaluated in a timely manner to ensure that the enterprise always maintains its sensitivity and response capabilities to risks. The early warning system has a high degree of automation and can automatically run and issue alarms without human intervention. This automated feature reduces the burden of manual monitoring and improves the efficiency and accuracy of early warning. The early warning system can be customized and adjusted according to the actual needs of the enterprise. Enterprises can set different thresholds and alarm methods according to their own risk tolerance and management needs to meet personalized risk management needs.

[0112] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A bankruptcy data analysis method based on big data, characterized in that: The method comprises: S1: Obtain company financial data, production data, and market sentiment data, and combine the three data as the first data; S2: preprocessing the first data to obtain second data, and storing the second data in a database; S3: Divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain a sentiment evaluation coefficient trend table, insert financial data and production data into the table, and obtain a company operation situation trend table; S4: Conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessed value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

2. According to the bankruptcy data analysis method based on big data as described in claim 1, it is characterized in that: The S1 includes: S11: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss number in the distributed ledger through the API interface; S12: Obtain the production log of the production equipment, and collect the production data of the production equipment in real time through the production log; S13: Use crawler systems to collect relevant text data on social platforms as a data source for market sentiment analysis; S14: Merge the financial data acquired in S11, the production data acquired in S12, and the market sentiment data acquired in S13, and use the merged data as the first data.

3. According to the bankruptcy data analysis method based on big data as described in claim 1, it is characterized in that: The S2 includes: S21: After cleaning the first data, standardize the first data; S22: Construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: a generator learns the distribution of real data and generates simulated data similar to the real data. A judge then judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in a database.

4. According to the bankruptcy data analysis method based on big data as described in claim 1, it is characterized in that: The S3 includes: S31: setting emotion indicators, which include positive emotion indicators and negative emotion indicators, calculating emotion evaluation coefficients on a daily basis, and drawing an emotion evaluation coefficient trend chart based on the emotion evaluation coefficients; S32: Insert the financial data and operation data in the second data into the emotion evaluation coefficient trend table, and align the time to ensure that the emotion evaluation coefficient and the financial data and production data are aligned at the same time to obtain the company's operation status trend table.

5. According to the bankruptcy data analysis method based on big data as described in claim 1, it is characterized in that: The S4 includes: S41: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results; S42: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold value. If the assessment value corresponding to the risk assessment result is higher than the threshold value, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

6. A bankruptcy data analysis system based on big data, characterized in that: The system comprises: Data collection system: obtain company financial data, production data and market sentiment data, and combine the three data as the first data; Data preprocessing system: preprocessing the first data to obtain second data, and storing the second data in a database; Data analysis system: divide the market sentiment data in the second data into positive sentiment and negative sentiment, calculate the sentiment evaluation coefficient according to the number of days, obtain the sentiment evaluation coefficient trend table, insert the financial data and production data into the table, and obtain the company's operating situation trend table; Risk assessment system: conduct bankruptcy analysis on the company's operating trend chart, and conduct quantitative assessment of the company's bankruptcy risk based on the analysis results. When the assessment value of the bankruptcy risk exceeds the threshold, the early warning device will issue a warning.

7. The bankruptcy data analysis system based on big data according to claim 6, characterized in that: The data acquisition system comprises: Financial information extraction system: Create a blockchain platform for financial data, set up nodes on the blockchain platform, and extract the date and the corresponding profit and loss in the distributed ledger through the API interface; Production information extraction system: obtain the production log of production equipment and collect the production data of production equipment in real time through the production log; Crawler system: Use the crawler system to collect relevant text data on social platforms as a data source for market sentiment analysis; Data merging system: merges the financial data acquired by S11, the production data acquired by S12, and the market sentiment data acquired by S13, and uses the merged data as the first data.

8. The bankruptcy data analysis system based on big data according to claim 6, characterized in that: The data preprocessing system comprises: Data standardization system: after cleaning the first data, standardize the first data; Simulated data training system: construct an adversarial network model and train the adversarial network model. The adversarial network model training includes: the generator learns the distribution of real data and generates simulated data similar to the real data. The judge then judges the real data and the simulated data. The generator and the judge are trained alternately. After the training is completed, the simulated data is added to the first data to obtain the second data, and the second data is stored in the database.

9. The bankruptcy data analysis system based on big data according to claim 6, characterized in that: The data analysis system comprises: Emotional scoring system: setting emotional indicators, including positive emotional indicators and negative emotional indicators, calculating emotional evaluation coefficients on a daily basis, and drawing an emotional evaluation coefficient trend chart based on the emotional evaluation coefficients; Time synchronization system: insert the financial data and operation data in the second data into the sentiment evaluation coefficient trend table, and align the time to ensure that the sentiment evaluation coefficient, financial data and production data are aligned at the same time to obtain the company's operation status trend table.

10. The bankruptcy data analysis system based on big data according to claim 6, characterized in that: The risk assessment system comprises: Risk assessment model: Establish a bankruptcy analysis model, analyze the company's operating trend chart, and conduct a quantitative assessment of the company's bankruptcy risk based on the analysis results; Early warning system: Compare the assessment value corresponding to the bankruptcy risk assessment result with the set threshold. If the assessment value corresponding to the risk assessment result is higher than the threshold, the early warning device will sound an alarm to warn the enterprise that there is a bankruptcy risk.

Images