Compliance inspection system and method for enterprise business process

Abstract

The invention relates to the technical field of business process compliance inspection systems, in particular to a compliance inspection system and method for an enterprise business process, and aims to solve the problems that in the prior art, an enterprise business process compliance inspection system generally adopts a static risk assessment model and lacks automatic and intelligent data acquisition and processing means, so that the enterprise business process compliance inspection system is inconvenient to use. And the accuracy and timeliness of risk assessment are insufficient. The system comprises a data acquisition module, a risk assessment module, a monitoring and early warning module, a map construction module, an intelligent recommendation module and a report generation module. According to the invention, through automatic data acquisition, preprocessing, risk assessment and monitoring early warning, the efficiency and accuracy of compliance inspection are significantly improved; by utilizing artificial intelligence and big data technologies, the system can quickly identify potential compliance risk points and perform quantitative evaluation, so that errors caused by manual intervention and subjective judgment are reduced, and the efficiency and accuracy of compliance inspection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of business process compliance checking systems, and in particular to a compliance checking system and method for enterprise business processes. Background Art

[0002] With the expansion of enterprise scale and the complexity of business, compliance management has become an indispensable part of enterprise operations; although the existing enterprise business process compliance inspection system can assist enterprises in compliance inspection to a certain extent, there are still some problems; for example, most of the existing systems rely on manual review, which is not only inefficient but also prone to errors, and it is difficult to meet the needs of modern enterprises to respond quickly to regulatory changes; in addition, the existing system also has deficiencies in data integration, risk identification and early warning, which makes it difficult for enterprises to timely discover and deal with potential compliance risks; specifically, the existing system often relies on manual review, which is not only inefficient but also easily affected by human factors, resulting in inconsistency and inaccuracy in audit results; there are often multiple information systems with different sources and different architectures within the enterprise, and the existing system has difficulties in data integration, making it difficult to achieve unified management and analysis; the existing system has limited capabilities in risk identification and early warning, making it difficult to timely discover and deal with potential compliance risks, resulting in enterprises facing higher legal risks. Summary of the invention

[0003] In order to overcome the problem that the enterprise business process compliance inspection system under the existing technology usually adopts a static risk assessment model and lacks automated and intelligent data collection and processing methods, resulting in insufficient accuracy and timeliness of risk assessment.

[0004] The technical solution of the present invention is: a compliance inspection system for enterprise business processes, comprising:

[0005] The data collection module is used to collect various types of data generated in the enterprise's business processes and perform data cleaning and data conversion on the collected data;

[0006] The risk assessment module is used to mine and analyze pre-processed data using artificial intelligence and big data technologies, establish an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conduct quantitative assessments on them;

[0007] The monitoring and early warning module is used to monitor business activities in real time, trigger the early warning mechanism when violations and potential risks are detected, and notify relevant personnel to handle them in a timely manner;

[0008] The graph construction module is used to integrate the internal and external compliance knowledge resources of the enterprise and build a compliance knowledge graph;

[0009] Intelligent recommendation module, which is used to recommend relevant compliance knowledge, training courses and operation guides to employees based on the compliance knowledge graph and employees’ compliance behavior data, to help employees improve their compliance awareness and capabilities;

[0010] The report generation module is used to generate compliance inspection reports based on the compliance inspection results and data analysis.

[0011] Preferably, various types of data generated in the business process of the enterprise are collected through the data collection module, and the collected data are cleaned and converted; the risk assessment module uses artificial intelligence and big data technology to mine and analyze the pre-processed data, establish an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conduct quantitative assessments on them; the monitoring and early warning module monitors business activities in real time, triggers the early warning mechanism when violations and potential risks are detected, and notifies relevant personnel to deal with them in a timely manner; the graph construction module integrates internal and external compliance knowledge resources of the enterprise to build a compliance knowledge graph; the intelligent recommendation module recommends relevant compliance knowledge, training courses and operating guides to employees based on the compliance knowledge graph and employees' compliance behavior data, to help employees improve their compliance awareness and capabilities; the report generation module generates a compliance inspection report based on the results of the compliance inspection and data analysis.

[0012] Preferably, the data collection module includes a data collection unit and a data preprocessing unit; the data collection unit is used to collect transaction records, employee behavior logs, and external regulatory policy data through crawler technology and data collection, and the data preprocessing module is used to clean the data, remove redundant and erroneous information, ensure the neatness of the data, and convert the data into a unified format through format conversion to facilitate subsequent processing and analysis; when the data collection module is working, the specific steps are as follows:

[0013] S201: Determine the types of data that need to be collected in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies;

[0014] S202: Identify and enumerate all relevant data sources, including internal enterprise systems and external data sources;

[0015] S203: For each data source, select the corresponding data collection method, using API interface and database query method for internal system, and using crawler technology and purchasing data service method for external data source;

[0016] S204: Implement the data collection plan and set a scheduled collection task for data with high real-time requirements;

[0017] S205: Conduct a preliminary review of the collected data to identify and process missing values, outliers, and duplicate data;

[0018] S206: Perform data cleaning operations according to data quality and business requirements, including format conversion and data standardization;

[0019] S207: Apply a data deduplication algorithm to perform deduplication processing on the cleaned data;

[0020] S208: According to the subsequent data processing and analysis requirements, the data is converted into a unified format, including conversion from text format to numerical format and unification of different time formats;

[0021] S209: Storing the preprocessed data in a database.

[0022] As a preferred option, the risk assessment module uses artificial intelligence and big data technology to mine and analyze pre-processed data, establishes an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conducts quantitative assessments to determine risk levels and potential impacts. Based on the assessment results, it provides enterprises with accurate risk warnings and prevention and control suggestions to help them respond to compliance risks in a timely manner. The risk assessment model uses a neural network machine learning algorithm to train and learn historical data to improve the accuracy and predictive ability of the model. At the same time, it uses natural language processing technology to perform semantic analysis and understanding of laws and regulations to ensure the accuracy and compliance of risk assessments. When the risk assessment module is working, the specific steps are as follows:

[0023] S301: Acquire cleaned business data in a unified format from a data collection module;

[0024] S302: further extracting and selecting features from the data to construct an input feature vector for the neural network model;

[0025] S303: Select a convolutional neural network architecture, define the number of layers of the neural network, the number of neurons in each layer, the activation function, and the loss function;

[0026] S304: Divide the prepared data set into a training set and a validation set;

[0027] S305: Use the training set to train the neural network model, adjust the model parameters through the back propagation algorithm, and minimize the loss function;

[0028] S306: Use the validation set to validate the model to prevent overfitting, and adjust the model parameters and architecture according to the validation results;

[0029] S307: After the training is completed, the new business data is input into the neural network model for compliance risk assessment;

[0030] S308: The model outputs risk scores and risk levels to identify potential compliance risk points;

[0031] S309: Obtain relevant laws, regulations, policies, and industry standard text data from the data collection module;

[0032] S310: Perform pre-processing operations such as cleaning, word segmentation, and stop word removal on the text data to construct the input of the natural language processing model;

[0033] S311: Use word embedding technology Word2Vec to convert text data into vector representation;

[0034] S312: Extract key information and compliance requirements from text data using semantic similarity calculation and entity recognition natural language processing technology;

[0035] S313: Match the extracted compliance requirements with business data to identify potential compliance risk points;

[0036] S314: Based on the matching results, quantitatively evaluate the risk points to determine the risk level and potential impact;

[0037] S315: Integrate and output the results of natural language processing and the risk assessment results of the neural network model.

[0038] As a preferred option, the monitoring and early warning module is integrated into the enterprise business process to achieve real-time monitoring of business activities. By setting early warning rules and thresholds, including specific transaction amounts and employee behavior patterns, once violations and potential risks are detected, the early warning mechanism is immediately triggered. The early warning methods include emails, text messages, and system notifications to ensure that relevant personnel can receive and process early warning information in a timely manner. The monitoring and early warning module uses stream processing technology to process and analyze real-time data. The early warning mechanism combines the rule engine and event trigger mechanism to achieve fast and accurate early warning response; when the monitoring and early warning module is working, the specific steps are as follows:

[0039] S401: Read monitoring rules, warning thresholds, and warning mode parameters from a configuration file;

[0040] S402: Establishing a connection with a data acquisition module to obtain real-time business data;

[0041] S403: Setting warning trigger conditions, including email sending and SMS notification;

[0042] S404: Obtain data in the enterprise business process in real time through the data interface;

[0043] S405: Cleaning and format conversion processing operations are performed on the acquired data, and key features in the data are extracted according to monitoring rules;

[0044] S406: Match the extracted features with the warning rules to determine whether there are any violations or potential risks;

[0045] S407: If a risk is detected, a risk assessment model is used to quantitatively assess the risk and determine the risk level and potential impact;

[0046] S408: Determine whether to trigger the early warning mechanism based on the risk assessment results;

[0047] S409: If an early warning needs to be triggered, send early warning information to relevant personnel according to the preset early warning method, including risk type, level, and recommended handling measures;

[0048] S410: After receiving the warning information, relevant personnel shall respond to the risk according to the recommended treatment measures;

[0049] S411: During the entire monitoring process, the monitoring log is recorded in real time, including data acquisition time, preprocessing results, risk detection results, and warning triggering conditions.

[0050] As a preferred option, the graph construction module integrates the internal and external compliance knowledge resources of the enterprise, including laws and regulations, industry standards, and internal rules and regulations of the enterprise. It uses the graph database and graph construction tool Neo4j to build a knowledge graph. By defining the attributes of nodes and edges, as well as the relationship types between nodes, it realizes the structured and visual display of compliance knowledge. By using the knowledge graph technology, it builds a compliance knowledge graph to display the association between different compliance requirements. The nodes in the graph represent the compliance requirements, and the edges represent the relationship between the requirements, so that the enterprise can better understand and apply compliance knowledge. When the graph construction module is working, the specific steps are as follows:

[0051] S501: Determine the scope of data collection, including laws and regulations, industry standards, and internal corporate rules and regulations, collect data, and conduct preliminary screening of the data to remove irrelevant or duplicate information;

[0052] S502: Cleaning the collected data, including removing spaces and special characters, and correcting typos;

[0053] S503: Perform word segmentation processing on the text data by using a rule-based word segmentation method, extract keywords, and use regular expressions to unify the data format;

[0054] S504: Determine the nodes and edges of the graph, as well as the attributes, and define the relationship type between the nodes;

[0055] S505: Create a graph using the graph database Neo4j's graph representation method based on the adjacency matrix;

[0056] S506: Import the preprocessed data into the graph database, and create nodes and edges according to the defined graph structure;

[0057] S507: Use the API of the graph database to set node and edge attributes, and add attributes to the nodes and edges, including name, description, and weight;

[0058] S508: constructing a query interface by using a query method based on depth-first search and a method for keyword retrieval using a string matching algorithm KMP, allowing users to query the graph by keyword and node type conditions;

[0059] S509: Use the visualization tool Graphviz to visualize the graph, and use the force-directed algorithm to determine the location of the nodes and the direction of the edges.

[0060] As a preferred option, the intelligent recommendation module adopts collaborative filtering algorithm and content-based recommendation algorithm, combined with employees' personal information and behavioral data to achieve personalized compliance knowledge recommendations. At the same time, it uses the association relationships in the knowledge graph to provide employees with a comprehensive compliance knowledge system and learning path. According to personalized information such as employees' positions and business operations, it recommends relevant compliance knowledge, training courses and operation guides. Through intelligent recommendations, it helps employees improve their compliance awareness and capabilities and reduce compliance risks.

[0061] As a preferred embodiment, when the intelligent recommendation module is working, the specific steps are as follows:

[0062] S701: Obtaining employees’ personal information from the company’s internal system, including position, rank, and department;

[0063] S702: Collect employees’ compliance behavior data in real time through the monitoring and early warning module, including operation records and transaction records;

[0064] S703: Extract features from employee information and compliance behavior data, including job characteristics and operating habits;

[0065] S704: Select an appropriate recommendation algorithm based on the features; recommend similar compliance knowledge based on the employee's historical behavior data through a collaborative filtering algorithm; recommend relevant compliance knowledge based on the employee's current operations and needs through a content-based recommendation algorithm;

[0066] S705: Using historical data to train and optimize the recommendation algorithm;

[0067] S706: Generate recommendation results using the trained recommendation algorithm based on the employee’s personal information and current behavior, including relevant compliance knowledge, training courses, and operation guides;

[0068] S707: Present the recommendation results to employees in the form of lists and charts.

[0069] As a preferred option, the report generation module generates a compliance inspection report based on the results of the compliance inspection and data analysis. The report is rich in content, including compliance risk points, rectification suggestions, and compliance trends; the report content is presented in the form of charts and dashboards; the report generation uses report generation tools and visualization technology to define report templates and styles, as well as the method and logic of data display, to achieve automatic generation and visualization of compliance inspection reports, and at the same time, combined with data analysis tools, in-depth mining and analysis of compliance data to provide accurate and comprehensive data support for the report.

[0070] Preferably, when the report generation module is working, the specific steps are as follows:

[0071] S901: Collect relevant compliance data from each module and integrate and store the data from each module using the data interface;

[0072] S902: Clean, convert and normalize the collected relevant data;

[0073] S903: Use decision tree machine learning algorithms to analyze data and identify potential compliance risk points;

[0074] S904: Assess risk levels and predict compliance trends based on historical data and business rules;

[0075] S905: Select a report template based on the compliance analysis results and the needs of the enterprise. The template includes the basic structure, style and presentation of the report;

[0076] S906: Use the report generation tool Tableau to fill the analyzed data into the report template based on the compliance analysis results to generate specific report content, including detailed description of compliance risk points, rectification suggestions, and compliance trend analysis;

[0077] S907: Use visualization technology ECharts to display the data in the report in the form of charts and dashboards;

[0078] S908: Generate a complete compliance check report and allow users to export the report to PDF or Excel format.

[0079] A compliance checking method for an enterprise business process includes the following steps:

[0080] S1: Identify key data types in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies;

[0081] S2: Identify and access all data sources, and collect data using API interfaces, database queries, crawler technology, and data services;

[0082] S3: Clean the collected data, remove redundant and erroneous information, and ensure the neatness of the data;

[0083] S4: Perform data conversion, including format unification, data standardization and deduplication, to facilitate subsequent analysis;

[0084] S5: Extract features from the preprocessed data and construct the input feature vector of the neural network model;

[0085] S6: Use convolutional neural network architecture to train risk assessment models and learn and optimize model parameters through historical data;

[0086] S7: Conduct risk assessment on new business data, output risk scores and levels, and identify potential compliance risk points;

[0087] S8: Use natural language processing technology to perform semantic analysis on regulations and policies, extract compliance requirements and match them with business data to further verify risk points;

[0088] S9: Configure monitoring rules, warning thresholds, and warning methods, such as email, SMS, and system notifications;

[0089] S10: Acquire data from enterprise business processes in real time, and perform cleaning and format conversion;

[0090] S11: Analyze the data according to the monitoring rules to determine whether there are any violations or potential risks;

[0091] S12: Once a risk is detected, the early warning mechanism is triggered immediately and early warning information is sent to relevant personnel, including the risk type, level and recommended treatment measures;

[0092] S13: Collect and screen internal and external compliance knowledge resources, such as laws and regulations, industry standards, and internal corporate rules and regulations;

[0093] S14: Clean and segment the collected data, extract keywords and unify the data format;

[0094] S15: Use the graph database Neo4j to build a compliance knowledge graph, define nodes, edges, and attributes, and display the relationship between compliance requirements;

[0095] S16: Obtain employees’ personal information and compliance behavior data, extract features to select appropriate recommendation algorithms;

[0096] S17: Use collaborative filtering and content-based recommendation algorithms, combined with the associations in the knowledge graph, to recommend relevant compliance knowledge, training courses and operation guides to employees;

[0097] S18: Present the recommendation results to employees in the form of lists and charts to help them improve their compliance awareness and capabilities;

[0098] S19: Collect compliance data from each module and perform cleaning, format conversion and normalization;

[0099] S20: Use machine learning algorithms such as decision trees to analyze data, identify compliance risk points and assess risk levels;

[0100] S21: Select a report template based on the analysis results and enterprise needs, and use the report generation tool to fill the analyzed data into the template;

[0101] S22: Use visualization technology to display the data in the report in the form of charts and dashboards to generate a complete compliance inspection report;

[0102] S23: Allows users to export reports to PDF, Excel and other formats for easy sharing and archiving;

[0103] S24: Regularly collect user feedback on the system, including report accuracy, recommendation effectiveness, and early warning response;

[0104] S25: Optimize and iterate the system based on feedback, including updating the risk assessment model, improving the knowledge graph, and optimizing the recommendation algorithm.

[0105] Beneficial effects of the present invention:

[0106] 1. Compared with the existing technology, the enterprise business process compliance inspection system usually adopts a static risk assessment model and lacks automated and intelligent data collection and processing methods, resulting in insufficient accuracy and timeliness of risk assessment. This invention significantly improves the efficiency and accuracy of compliance inspection through automated data collection, preprocessing, risk assessment and monitoring and early warning. By using artificial intelligence and big data technology, the system can quickly identify potential compliance risk points and conduct quantitative assessments, reducing errors caused by manual intervention and subjective judgment, and improving the efficiency and accuracy of compliance inspections.

[0107] 2. Through the risk assessment module, artificial intelligence and big data technologies are used to mine and analyze pre-processed data. By establishing an intelligent compliance risk assessment model, potential compliance risk points are automatically identified and quantitatively assessed to determine risk levels and potential impacts. Based on the assessment results, accurate risk warnings and prevention and control suggestions are provided to enterprises to help them respond to compliance risks in a timely manner. The risk assessment model uses a neural network machine learning algorithm to train and learn historical data to improve the accuracy and predictive ability of the model. At the same time, natural language processing technology is used to perform semantic analysis and understanding of laws and regulations to ensure the accuracy and compliance of risk assessments; thereby reducing errors caused by manual intervention and subjective judgment, and improving the efficiency and accuracy of compliance inspections;

[0108] 3. By integrating the monitoring and early warning module into the enterprise business process, real-time monitoring of business activities can be achieved. By setting early warning rules and thresholds, including specific transaction amounts and employee behavior patterns, once violations and potential risks are detected, the early warning mechanism is immediately triggered. The early warning methods include emails, text messages, and system notifications to ensure that relevant personnel can receive and handle the early warning information in a timely manner. The monitoring and early warning module uses stream processing technology to process and analyze real-time data. The early warning mechanism combines the rule engine and event trigger mechanism to achieve fast and accurate early warning response. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 What is shown is a schematic diagram of the architecture of the compliance inspection system for the enterprise business process of the present invention;

[0110] Figure 2 What is shown is a workflow diagram of a data collection module of a compliance inspection system for an enterprise business process of the present invention;

[0111] Figure 3 What is shown is a schematic diagram of the workflow of the risk assessment module of the compliance inspection system for the enterprise business process of the present invention;

[0112] Figure 4 What is shown is a workflow diagram of a monitoring and early warning module of a compliance inspection system for an enterprise business process of the present invention;

[0113] Figure 5 What is shown is a schematic diagram of the workflow of the enterprise business process compliance inspection system atlas construction module of the present invention;

[0114] Figure 6 What is shown is a schematic diagram of the workflow of the intelligent recommendation module of the compliance inspection system for the enterprise business process of the present invention;

[0115] Figure 7 What is shown is a workflow diagram of a report generation module of a compliance inspection system for an enterprise business process of the present invention;

[0116] Figure 8 What is shown is a flowchart of the steps of the compliance inspection method for the enterprise business process of the present invention;

[0117] Fig. 9 What is presented is a workflow diagram of a compliance checking system for enterprise business processes of the present invention. DETAILED DESCRIPTION

[0118] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0119] See also Figure 1-9 The present invention provides an embodiment: a compliance inspection system for an enterprise business process, comprising:

[0120] The data collection module is used to collect various types of data generated in the enterprise's business processes and perform data cleaning and data conversion on the collected data;

[0121] The risk assessment module is used to mine and analyze pre-processed data using artificial intelligence and big data technologies, establish an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conduct quantitative assessments on them;

[0122] The monitoring and early warning module is used to monitor business activities in real time, trigger the early warning mechanism when violations and potential risks are detected, and notify relevant personnel to handle them in a timely manner;

[0123] The graph construction module is used to integrate the internal and external compliance knowledge resources of the enterprise and build a compliance knowledge graph;

[0124] Intelligent recommendation module, which is used to recommend relevant compliance knowledge, training courses and operation guides to employees based on the compliance knowledge graph and employees’ compliance behavior data, to help employees improve their compliance awareness and capabilities;

[0125] The report generation module is used to generate compliance inspection reports based on the compliance inspection results and data analysis.

[0126] Preferably, various types of data generated in the business process of the enterprise are collected through the data collection module, and the collected data are cleaned and converted; the risk assessment module uses artificial intelligence and big data technology to mine and analyze the pre-processed data, establish an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conduct quantitative assessments on them; the monitoring and early warning module monitors business activities in real time, triggers the early warning mechanism when violations and potential risks are detected, and notifies relevant personnel to deal with them in a timely manner; the graph construction module integrates internal and external compliance knowledge resources of the enterprise to build a compliance knowledge graph; the intelligent recommendation module recommends relevant compliance knowledge, training courses and operating guides to employees based on the compliance knowledge graph and employees' compliance behavior data, to help employees improve their compliance awareness and capabilities; the report generation module generates a compliance inspection report based on the results of the compliance inspection and data analysis.

[0127] Preferably, the data collection module includes a data collection unit and a data preprocessing unit; the data collection unit is used to collect transaction records, employee behavior logs, and external regulatory policy data through crawler technology and data collection, and the data preprocessing module is used to clean the data, remove redundant and erroneous information, ensure the neatness of the data, and convert the data into a unified format through format conversion to facilitate subsequent processing and analysis; when the data collection module is working, the specific steps are as follows:

[0128] S201: Determine the types of data that need to be collected in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies;

[0129] S202: Identify and enumerate all relevant data sources, including internal enterprise systems and external data sources;

[0130] S203: For each data source, select the corresponding data collection method, using API interface and database query method for internal system, and using crawler technology and purchasing data service method for external data source;

[0131] S204: Implement the data collection plan and set a scheduled collection task for data with high real-time requirements;

[0132] S205: Conduct a preliminary review of the collected data to identify and process missing values, outliers, and duplicate data;

[0133] S206: Perform data cleaning operations according to data quality and business requirements, including format conversion and data standardization;

[0134] S207: Apply a data deduplication algorithm to perform deduplication processing on the cleaned data;

[0135] S208: According to the subsequent data processing and analysis requirements, the data is converted into a unified format, including conversion from text format to numerical format and unification of different time formats;

[0136] S209: Storing the preprocessed data in a database.

[0137] As a preferred option, the risk assessment module uses artificial intelligence and big data technology to mine and analyze pre-processed data, establishes an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conducts quantitative assessments to determine risk levels and potential impacts. Based on the assessment results, it provides enterprises with accurate risk warnings and prevention and control suggestions to help them respond to compliance risks in a timely manner. The risk assessment model uses a neural network machine learning algorithm to train and learn historical data to improve the accuracy and predictive ability of the model. At the same time, it uses natural language processing technology to perform semantic analysis and understanding of laws and regulations to ensure the accuracy and compliance of risk assessments. When the risk assessment module is working, the specific steps are as follows:

[0138] S301: Acquire cleaned business data in a unified format from a data collection module;

[0139] S302: further extracting and selecting features from the data to construct an input feature vector for the neural network model;

[0140] S303: Select a convolutional neural network architecture, define the number of layers of the neural network, the number of neurons in each layer, the activation function, and the loss function;

[0141] S304: Divide the prepared data set into a training set and a validation set;

[0142] S305: Use the training set to train the neural network model, adjust the model parameters through the back propagation algorithm, and minimize the loss function;

[0143] S306: Use the validation set to validate the model to prevent overfitting, and adjust the model parameters and architecture according to the validation results;

[0144] S307: After the training is completed, the new business data is input into the neural network model for compliance risk assessment;

[0145] S308: The model outputs risk scores and risk levels to identify potential compliance risk points;

[0146] S309: Obtain relevant laws, regulations, policies, and industry standard text data from the data collection module;

[0147] S310: Perform pre-processing operations such as cleaning, word segmentation, and stop word removal on the text data to construct the input of the natural language processing model;

[0148] S311: Use word embedding technology Word2Vec to convert text data into vector representation;

[0149] S312: Extract key information and compliance requirements from text data using semantic similarity calculation and entity recognition natural language processing technology;

[0150] S313: Match the extracted compliance requirements with business data to identify potential compliance risk points;

[0151] S314: Based on the matching results, quantitatively evaluate the risk points to determine the risk level and potential impact;

[0152] S315: Integrate and output the results of natural language processing and the risk assessment results of the neural network model.

[0153] As a preferred option, the monitoring and early warning module is integrated into the enterprise business process to achieve real-time monitoring of business activities. By setting early warning rules and thresholds, including specific transaction amounts and employee behavior patterns, once violations and potential risks are detected, the early warning mechanism is immediately triggered. The early warning methods include emails, text messages, and system notifications to ensure that relevant personnel can receive and process early warning information in a timely manner. The monitoring and early warning module uses stream processing technology to process and analyze real-time data. The early warning mechanism combines the rule engine and event trigger mechanism to achieve fast and accurate early warning response; when the monitoring and early warning module is working, the specific steps are as follows:

[0154] S401: Read monitoring rules, warning thresholds, and warning mode parameters from a configuration file;

[0155] S402: Establishing a connection with a data acquisition module to obtain real-time business data;

[0156] S403: Setting warning trigger conditions, including email sending and SMS notification;

[0157] S404: Obtain data in the enterprise business process in real time through the data interface;

[0158] S405: Cleaning and format conversion processing operations are performed on the acquired data, and key features in the data are extracted according to monitoring rules;

[0159] S406: Match the extracted features with the warning rules to determine whether there are any violations or potential risks;

[0160] S407: If a risk is detected, a risk assessment model is used to quantitatively assess the risk and determine the risk level and potential impact;

[0161] S408: Determine whether to trigger the early warning mechanism based on the risk assessment results;

[0162] S409: If an early warning needs to be triggered, send early warning information to relevant personnel according to the preset early warning method, including risk type, level, and recommended handling measures;

[0163] S410: After receiving the warning information, relevant personnel shall respond to the risk according to the recommended treatment measures;

[0164] S411: During the entire monitoring process, the monitoring log is recorded in real time, including data acquisition time, preprocessing results, risk detection results, and warning triggering conditions.

[0165] As a preferred option, the graph construction module integrates the internal and external compliance knowledge resources of the enterprise, including laws and regulations, industry standards, and internal rules and regulations of the enterprise. It uses the graph database and graph construction tool Neo4j to build a knowledge graph. By defining the attributes of nodes and edges, as well as the relationship types between nodes, it realizes the structured and visual display of compliance knowledge. By using the knowledge graph technology, it builds a compliance knowledge graph to display the association between different compliance requirements. The nodes in the graph represent the compliance requirements, and the edges represent the relationship between the requirements, so that the enterprise can better understand and apply compliance knowledge. When the graph construction module is working, the specific steps are as follows:

[0166] S501: Determine the scope of data collection, including laws and regulations, industry standards, and internal corporate rules and regulations, collect data, and conduct preliminary screening of the data to remove irrelevant or duplicate information;

[0167] S502: Cleaning the collected data, including removing spaces and special characters, and correcting typos;

[0168] S503: Perform word segmentation processing on the text data by using a rule-based word segmentation method, extract keywords, and use regular expressions to unify the data format;

[0169] S504: Determine the nodes and edges of the graph, as well as the attributes, and define the relationship type between the nodes;

[0170] S505: Create a graph using the graph database Neo4j's graph representation method based on the adjacency matrix;

[0171] S506: Import the preprocessed data into the graph database, and create nodes and edges according to the defined graph structure;

[0172] S507: Use the API of the graph database to set node and edge attributes, and add attributes to the nodes and edges, including name, description, and weight;

[0173] S508: constructing a query interface by using a query method based on depth-first search and a method for keyword retrieval using a string matching algorithm KMP, allowing users to query the graph by keyword and node type conditions;

[0174] S509: Use the visualization tool Graphviz to visualize the graph, and use the force-directed algorithm to determine the location of the nodes and the direction of the edges.

[0175] As a preferred option, the intelligent recommendation module adopts collaborative filtering algorithm and content-based recommendation algorithm, combined with employees' personal information and behavioral data to achieve personalized compliance knowledge recommendations. At the same time, it uses the association relationships in the knowledge graph to provide employees with a comprehensive compliance knowledge system and learning path. According to personalized information such as employees' positions and business operations, it recommends relevant compliance knowledge, training courses and operation guides. Through intelligent recommendations, it helps employees improve their compliance awareness and capabilities and reduce compliance risks.

[0176] As a preferred embodiment, when the intelligent recommendation module is working, the specific steps are as follows:

[0177] S701: Obtaining employees’ personal information from the company’s internal system, including position, rank, and department;

[0178] S702: Collect employees’ compliance behavior data in real time through the monitoring and early warning module, including operation records and transaction records;

[0179] S703: Extract features from employee information and compliance behavior data, including job characteristics and operating habits;

[0180] S704: Select an appropriate recommendation algorithm based on the features; recommend similar compliance knowledge based on the employee's historical behavior data through a collaborative filtering algorithm; recommend relevant compliance knowledge based on the employee's current operations and needs through a content-based recommendation algorithm;

[0181] S705: Using historical data to train and optimize the recommendation algorithm;

[0182] S706: Generate recommendation results using the trained recommendation algorithm based on the employee’s personal information and current behavior, including relevant compliance knowledge, training courses, and operation guides;

[0183] S707: Present the recommendation results to employees in the form of lists and charts.

[0184] As a preferred option, the report generation module generates a compliance inspection report based on the results of the compliance inspection and data analysis. The report is rich in content, including compliance risk points, rectification suggestions, and compliance trends; the report content is presented in the form of charts and dashboards; the report generation uses report generation tools and visualization technology to define report templates and styles, as well as the method and logic of data display, to achieve automatic generation and visualization of compliance inspection reports, and at the same time, combined with data analysis tools, in-depth mining and analysis of compliance data to provide accurate and comprehensive data support for the report.

[0185] Preferably, when the report generation module is working, the specific steps are as follows:

[0186] S901: Collect relevant compliance data from each module and integrate and store the data from each module using the data interface;

[0187] S902: Clean, convert and normalize the collected relevant data;

[0188] S903: Use decision tree machine learning algorithms to analyze data and identify potential compliance risk points;

[0189] S904: Assess risk levels and predict compliance trends based on historical data and business rules;

[0190] S905: Select a report template based on the compliance analysis results and the needs of the enterprise. The template includes the basic structure, style and presentation of the report;

[0191] S906: Use the report generation tool Tableau to fill the analyzed data into the report template based on the compliance analysis results to generate specific report content, including detailed description of compliance risk points, rectification suggestions, and compliance trend analysis;

[0192] S907: Use visualization technology ECharts to display the data in the report in the form of charts and dashboards;

[0193] S908: Generate a complete compliance check report and allow users to export the report to PDF or Excel format.

[0194] A compliance checking method for an enterprise business process includes the following steps:

[0195] S1: Identify key data types in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies;

[0196] S2: Identify and access all data sources, and collect data using API interfaces, database queries, crawler technology, and data services;

[0197] S3: Clean the collected data, remove redundant and erroneous information, and ensure the neatness of the data;

[0198] S4: Perform data conversion, including format unification, data standardization and deduplication, to facilitate subsequent analysis;

[0199] S5: Extract features from the preprocessed data and construct the input feature vector of the neural network model;

[0200] S6: Use convolutional neural network architecture to train risk assessment models and learn and optimize model parameters through historical data;

[0201] S7: Conduct risk assessment on new business data, output risk scores and levels, and identify potential compliance risk points;

[0202] S8: Use natural language processing technology to perform semantic analysis on regulations and policies, extract compliance requirements and match them with business data to further verify risk points;

[0203] S9: Configure monitoring rules, warning thresholds, and warning methods, such as email, SMS, and system notifications;

[0204] S10: Acquire data from enterprise business processes in real time, and perform cleaning and format conversion;

[0205] S11: Analyze the data according to the monitoring rules to determine whether there are any violations or potential risks;

[0206] S12: Once a risk is detected, the early warning mechanism is triggered immediately and early warning information is sent to relevant personnel, including the risk type, level and recommended treatment measures;

[0207] S13: Collect and screen internal and external compliance knowledge resources, such as laws and regulations, industry standards, and internal corporate rules and regulations;

[0208] S14: Clean and segment the collected data, extract keywords and unify the data format;

[0209] S15: Use the graph database Neo4j to build a compliance knowledge graph, define nodes, edges, and attributes, and display the relationship between compliance requirements;

[0210] S16: Obtain employees’ personal information and compliance behavior data, extract features to select appropriate recommendation algorithms;

[0211] S17: Use collaborative filtering and content-based recommendation algorithms, combined with the associations in the knowledge graph, to recommend relevant compliance knowledge, training courses and operation guides to employees;

[0212] S18: Present the recommendation results to employees in the form of lists and charts to help them improve their compliance awareness and capabilities;

[0213] S19: Collect compliance data from each module and perform cleaning, format conversion and normalization;

[0214] S20: Use machine learning algorithms such as decision trees to analyze data, identify compliance risk points and assess risk levels;

[0215] S21: Select a report template based on the analysis results and enterprise needs, and use the report generation tool to fill the analyzed data into the template;

[0216] S22: Use visualization technology to display the data in the report in the form of charts and dashboards to generate a complete compliance inspection report;

[0217] S23: Allows users to export reports to PDF, Excel and other formats for easy sharing and archiving;

[0218] S24: Regularly collect user feedback on the system, including report accuracy, recommendation effectiveness, and early warning response;

[0219] S25: Optimize and iterate the system based on feedback, including updating the risk assessment model, improving the knowledge graph, and optimizing the recommendation algorithm.

[0220] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A compliance checking system for enterprise business processes; characterized by: Included are: The data collection module is used to collect various types of data generated in the enterprise's business processes and perform data cleaning and data conversion on the collected data; The risk assessment module is used to mine and analyze pre-processed data using artificial intelligence and big data technologies, establish an intelligent compliance risk assessment model to automatically identify potential compliance risk points, and conduct quantitative assessments on them; The monitoring and early warning module is used to monitor business activities in real time, trigger the early warning mechanism when violations and potential risks are detected, and notify relevant personnel to handle them in a timely manner; The graph construction module is used to integrate the internal and external compliance knowledge resources of the enterprise and build a compliance knowledge graph; Intelligent recommendation module, which is used to recommend relevant compliance knowledge, training courses and operation guides to employees based on the compliance knowledge graph and employees’ compliance behavior data, to help employees improve their compliance awareness and capabilities; The report generation module is used to generate compliance inspection reports based on the compliance inspection results and data analysis.

2. The compliance checking system for enterprise business processes according to claim 1, characterized in that: The data collection module includes a data collection unit and a data preprocessing unit; the data collection unit is used to collect transaction records, employee behavior logs, and external regulatory policy data through crawler technology and data collection, and the data preprocessing module is used to clean the data, remove redundant and erroneous information, ensure the neatness of the data, and convert the data into a unified format through format conversion to facilitate subsequent processing and analysis; when the data collection module is working, the specific steps are as follows: S201: Determine the types of data that need to be collected in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies; S202: Identify and enumerate all relevant data sources, including internal enterprise systems and external data sources; S203: For each data source, select the corresponding data collection method, using API interface and database query method for internal system, and using crawler technology and purchasing data service method for external data source; S204: Implement the data collection plan and set a scheduled collection task for data with high real-time requirements; S205: Conduct a preliminary review of the collected data to identify and process missing values, outliers, and duplicate data; S206: Perform data cleaning operations according to data quality and business requirements, including format conversion and data standardization; S207: Apply a data deduplication algorithm to perform deduplication processing on the cleaned data; S208: According to the subsequent data processing and analysis requirements, the data is converted into a unified format, including conversion from text format to numerical format and unification of different time formats; S209: Storing the preprocessed data in a database.

3. The compliance checking system for enterprise business processes according to claim 2, characterized in that: The risk assessment module uses artificial intelligence and big data technologies to mine and analyze pre-processed data. It establishes an intelligent compliance risk assessment model to automatically identify potential compliance risk points, conduct quantitative assessments, determine risk levels and potential impacts, and provide companies with accurate risk warnings and prevention and control recommendations based on the assessment results to help companies respond to compliance risks in a timely manner. The risk assessment model uses a neural network machine learning algorithm to train and learn historical data to improve the accuracy and predictive ability of the model. At the same time, it uses natural language processing technology to perform semantic analysis and understanding of laws and regulations to ensure the accuracy and compliance of risk assessments. When the risk assessment module is working, the specific steps are as follows: S301: Acquire cleaned business data in a unified format from a data collection module; S302: further extracting and selecting features from the data to construct an input feature vector for the neural network model; S303: Select a convolutional neural network architecture, define the number of layers of the neural network, the number of neurons in each layer, the activation function, and the loss function; S304: Divide the prepared data set into a training set and a validation set; S305: Use the training set to train the neural network model, adjust the model parameters through the back propagation algorithm, and minimize the loss function; S306: Use the validation set to validate the model to prevent overfitting, and adjust the model parameters and architecture according to the validation results; S307: After the training is completed, the new business data is input into the neural network model for compliance risk assessment; S308: The model outputs risk scores and risk levels to identify potential compliance risk points; S309: Obtain relevant laws, regulations, policies, and industry standard text data from the data collection module; S310: Perform pre-processing operations such as cleaning, word segmentation, and stop word removal on the text data to construct the input of the natural language processing model; S311: Use word embedding technology Word2Vec to convert text data into vector representation; S312: Extract key information and compliance requirements from text data using semantic similarity calculation and entity recognition natural language processing technology; S313: Match the extracted compliance requirements with business data to identify potential compliance risk points; S314: Based on the matching results, quantitatively evaluate the risk points to determine the risk level and potential impact; S315: Integrate and output the results of natural language processing and the risk assessment results of the neural network model.

4. The compliance checking system for enterprise business processes according to claim 3 is characterized by: The monitoring and early warning module is integrated into the enterprise business process to achieve real-time monitoring of business activities. By setting early warning rules and thresholds, including specific transaction amounts and employee behavior patterns, once violations and potential risks are detected, the early warning mechanism is immediately triggered. The early warning methods include emails, text messages, and system notifications to ensure that relevant personnel can receive and handle early warning information in a timely manner. The monitoring and early warning module uses stream processing technology to process and analyze real-time data. The early warning mechanism combines the rule engine and event trigger mechanism to achieve fast and accurate early warning response; when the monitoring and early warning module is working, the specific steps are as follows: S401: Read monitoring rules, warning thresholds, and warning mode parameters from a configuration file; S402: Establishing a connection with a data acquisition module to obtain real-time business data; S403: Setting warning trigger conditions, including email sending and SMS notification; S404: Obtain data in the enterprise business process in real time through the data interface; S405: Cleaning and format conversion processing operations are performed on the acquired data, and key features in the data are extracted according to monitoring rules; S406: Match the extracted features with the warning rules to determine whether there are any violations or potential risks; S407: If a risk is detected, a risk assessment model is used to quantitatively assess the risk and determine the risk level and potential impact; S408: Determine whether to trigger the early warning mechanism based on the risk assessment results; S409: If an early warning needs to be triggered, send early warning information to relevant personnel according to the preset early warning method, including risk type, level, and recommended handling measures; S410: After receiving the warning information, relevant personnel shall respond to the risk according to the recommended treatment measures; S411: During the entire monitoring process, the monitoring log is recorded in real time, including data acquisition time, preprocessing results, risk detection results, and warning triggering conditions.

5. The compliance checking system for enterprise business processes according to claim 4, characterized in that: The graph construction module integrates internal and external compliance knowledge resources of the enterprise, including laws and regulations, industry standards, and internal rules and regulations of the enterprise. It uses the graph database and graph construction tool Neo4j to build a knowledge graph. By defining the attributes of nodes and edges, as well as the relationship types between nodes, it realizes the structured and visual display of compliance knowledge. By using knowledge graph technology, it builds a compliance knowledge graph to display the association between different compliance requirements. The nodes in the graph represent compliance requirements, and the edges represent the relationship between requirements, so that enterprises can better understand and apply compliance knowledge. When the graph construction module is working, the specific steps are as follows: S501: Determine the scope of data collection, including laws and regulations, industry standards, and internal corporate rules and regulations, collect data, and conduct preliminary screening of the data to remove irrelevant or duplicate information; S502: Cleaning the collected data, including removing spaces and special characters, and correcting typos; S503: Perform word segmentation processing on the text data by using a rule-based word segmentation method, extract keywords, and use regular expressions to unify the data format; S504: Determine the nodes and edges of the graph, as well as the attributes, and define the relationship type between the nodes; S505: Create a graph using the graph database Neo4j's graph representation method based on the adjacency matrix; S506: Import the preprocessed data into the graph database, and create nodes and edges according to the defined graph structure; S507: Use the API of the graph database to set node and edge attributes, and add attributes to the nodes and edges, including name, description, and weight; S508: constructing a query interface by using a query method based on depth-first search and a method for keyword retrieval using a string matching algorithm KMP, allowing users to query the graph by keyword and node type conditions; S509: Use the visualization tool Graphviz to visualize the graph, and use the force-directed algorithm to determine the location of the nodes and the direction of the edges.

6. The compliance checking system for enterprise business processes according to claim 5, characterized in that: The intelligent recommendation module adopts collaborative filtering algorithm and content-based recommendation algorithm, combined with employees' personal information and behavioral data, to achieve personalized compliance knowledge recommendation. At the same time, it uses the association relationships in the knowledge graph to provide employees with a comprehensive compliance knowledge system and learning path. According to personalized information such as employees' positions and business operations, it recommends relevant compliance knowledge, training courses and operation guides. Through intelligent recommendations, it helps employees improve their compliance awareness and capabilities and reduce compliance risks.

7. The compliance checking system for enterprise business processes according to claim 6, characterized in that: When the intelligent recommendation module is working, the specific steps are as follows: S701: Obtaining employees’ personal information from the company’s internal system, including position, rank, and department; S702: Collect employees’ compliance behavior data in real time through the monitoring and early warning module, including operation records and transaction records; S703: Extract features from employee information and compliance behavior data, including job characteristics and operating habits; S704: Selecting a suitable recommendation algorithm according to the features; Recommend similar compliance knowledge based on employees’ historical behavior data through collaborative filtering algorithms; recommend relevant compliance knowledge based on employees’ current operations and needs through content-based recommendation algorithms; S705: Using historical data to train and optimize the recommendation algorithm; S706: Generate recommendation results using the trained recommendation algorithm based on the employee’s personal information and current behavior, including relevant compliance knowledge, training courses, and operation guides; S707: Present the recommendation results to employees in the form of lists and charts.

8. The enterprise business process compliance checking system according to claim 7, characterized in that: The report generation module generates compliance inspection reports based on the compliance inspection results and data analysis. The reports are rich in content, including compliance risk points, rectification suggestions, and compliance trends; Present report contents in the form of charts and dashboards; Report generation uses report generation tools and visualization technology to define report templates and styles, as well as the method and logic of data display, to achieve automatic generation and visual display of compliance inspection reports. At the same time, combined with data analysis tools, in-depth mining and analysis of compliance data is carried out to provide accurate and comprehensive data support for the report.

9. The enterprise business process compliance checking system according to claim 8, characterized in that: When the report generation module is working, the specific steps are as follows: S901: Collect relevant compliance data from each module and integrate and store the data from each module using the data interface; S902: Clean, convert and normalize the collected relevant data; S903: Use decision tree machine learning algorithms to analyze data and identify potential compliance risk points; S904: Assess risk levels and predict compliance trends based on historical data and business rules; S905: Select a report template based on the compliance analysis results and the needs of the enterprise. The template includes the basic structure, style and presentation of the report; S906: Use the report generation tool Tableau to fill the analyzed data into the report template based on the compliance analysis results to generate specific report content, including detailed description of compliance risk points, rectification suggestions, and compliance trend analysis; S907: Use visualization technology ECharts to display the data in the report in the form of charts and dashboards; S908: Generate a complete compliance check report and allow users to export the report to PDF or Excel format.

10. A compliance inspection method for an enterprise business process, characterized in that: The following steps are included: S1: Identify key data types in the enterprise's business processes, including transaction records, employee behavior logs, and external regulations and policies; S2: Identify and access all data sources, and collect data using API interfaces, database queries, crawler technology, and data services; S3: Clean the collected data, remove redundant and erroneous information, and ensure the neatness of the data; S4: Perform data conversion, including format unification, data standardization and deduplication, to facilitate subsequent analysis; S5: Extract features from the preprocessed data and construct the input feature vector of the neural network model; S6: Use convolutional neural network architecture to train risk assessment models and learn and optimize model parameters through historical data; S7: Conduct risk assessment on new business data, output risk scores and levels, and identify potential compliance risk points; S8: Use natural language processing technology to perform semantic analysis on regulations and policies, extract compliance requirements and match them with business data to further verify risk points; S9: Configure monitoring rules, warning thresholds, and warning methods, such as email, SMS, and system notifications; S10: Acquire data from enterprise business processes in real time, and perform cleaning and format conversion; S11: Analyze the data according to the monitoring rules to determine whether there are any violations or potential risks; S12: Once a risk is detected, the early warning mechanism is triggered immediately and early warning information is sent to relevant personnel, including the risk type, level and recommended treatment measures; S13: Collect and screen internal and external compliance knowledge resources, such as laws and regulations, industry standards, and internal corporate rules and regulations; S14: Clean and segment the collected data, extract keywords and unify the data format; S15: Use the graph database Neo4j to build a compliance knowledge graph, define nodes, edges, and attributes, and display the relationship between compliance requirements; S16: Obtain employees’ personal information and compliance behavior data, extract features to select appropriate recommendation algorithms; S17: Use collaborative filtering and content-based recommendation algorithms, combined with the associations in the knowledge graph, to recommend relevant compliance knowledge, training courses and operation guides to employees; S18: Present the recommendation results to employees in the form of lists and charts to help them improve their compliance awareness and capabilities; S19: Collect compliance data from each module and perform cleaning, format conversion and normalization; S20: Use machine learning algorithms such as decision trees to analyze data, identify compliance risk points and assess risk levels; S21: Select a report template based on the analysis results and enterprise needs, and use the report generation tool to fill the analyzed data into the template; S22: Use visualization technology to display the data in the report in the form of charts and dashboards to generate a complete compliance inspection report; S23: Allows users to export reports to PDF, Excel and other formats for easy sharing and archiving; S24: Regularly collect user feedback on the system, including report accuracy, recommendation effectiveness, and early warning response; S25: Optimize and iterate the system based on feedback, including updating the risk assessment model, improving the knowledge graph, and optimizing the recommendation algorithm.

Images