Investment and financing project archive evidence intelligent management platform based on big data

Abstract

The invention belongs to archives, and particularly relates to an investment and financing project archive evidence intelligent management platform based on big data, which comprises the following steps: step 1, a system uploads and preprocesses archives through an upload and storage module, and converts project files in different formats into structured data; 2, a data processing and intelligent analysis module automatically extracts key information and performs archive classification, data matching and compliance review by using natural language processing and a machine learning algorithm on the basis of data of the uploading and storage module; 3, the system carries out automatic examination based on the compliance model, an evidence chain meeting legal requirements is generated, real-time risk assessment is carried out, and when potential risks are recognized, the system triggers early warning and provides suggestions; step 4, in combination with risk early warning and legal examination data, a system workflow and task scheduling module dynamically schedules a workflow according to a task priority and a dependency relationship; and 5, providing decision support by the system through a user interface, and helping a manager to make a precise decision.

Description

Technical Field

[0001] The present invention relates to archives, and in particular to an intelligent management platform for archive evidence of investment and financing projects based on big data. Background Art

[0002] With the continuous expansion of the investment and financing market, especially in cross-industry and multi-field investment and financing projects, traditional file management methods face many challenges. The existing investment and financing file management systems are mostly based on manual review, with problems such as low processing efficiency, high manual error rate, and difficulty in adapting to rapidly changing laws and regulations. In particular, traditional systems are unable to provide intelligent support for large-scale data processing, compliance checks, and the generation of document evidence chains.

[0003] At present, most project archive management relies on simple data storage and query, and lacks automation functions based on artificial intelligence and data analysis. How to extract effective information from a large number of complex project documents and ensure data compliance and legal risk control has become a difficult problem that needs to be solved urgently in the industry.

[0004] In response to the above-mentioned technical defects, this paper proposes an intelligent management platform solution for investment and financing project archive evidence based on big data. Summary of the invention

[0005] To solve the above problems, the present invention provides the following technical solutions:

[0006] An intelligent management platform for investment and financing project archive evidence based on big data, including upload and storage modules, data processing and intelligent analysis modules, evidence chain generation modules, risk assessment and early warning modules, workflow and task scheduling modules, and user interaction and decision support modules;

[0007] The upload and storage module includes a data upload unit, a data storage unit and data backup and recovery. The data upload unit collects project files from different sources. After the files are uploaded to the system, they will first be pre-processed and then enter the data storage unit. The data storage unit uses a distributed database to store structured and unstructured data to ensure efficient access to massive data. At the same time, the data backup and recovery mechanism is used to ensure that data can be quickly restored when a failure occurs;

[0008] The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). The upload and storage module is processed based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, including file preprocessing and information extraction, file classification and matching and compliance review. The system automatically processes and analyzes the project files of the data collected and stored by the upload and storage module through multiple technical means such as file preprocessing and information extraction, file classification and matching and compliance review, thereby improving work efficiency and accuracy, and ensuring that the files comply with relevant laws and regulations.

[0009] The evidence chain generation module: based on the data processing and intelligent analysis module, uses deep learning (DL) and natural language processing (NLP) technology to automatically generate an evidence chain. Each key document or clause is automatically generated with an evidence chain associated with the document through file preprocessing and information extraction, file classification and data matching, and compliance review.

[0010] The risk assessment and early warning module: based on the evidence chain generation module and the data of historical projects, combined with the terms of the current contract, generates a risk score for the project and determines whether there are legal or financial risks;

[0011] The workflow and task scheduling module: based on the risk assessment and early warning module’s shared assessment and current workflow and tasks, adopts a rule-based task scheduling engine to automatically and dynamically adjust the task execution order and priority according to the project file content and progress;

[0012] In the user interaction and decision support module, users can access project archives through PC or mobile terminals, view archive processing results, compliance reports, risk assessment reports, and view system-recommended investment strategies or adjustment plans through decision support tools.

[0013] Further, the data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL), and the upload and storage module is processed based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, and the data upload collects different data PDF files, scanned archives or email multi-source files;

[0014] The data storage unit includes a distributed database to store structured and unstructured data. The distributed database (such as Hadoop HDFS, Cassandra) stores massive data to ensure data reliability and query speed. The unstructured data (such as scanned contract files, picture files, etc.) is stored in an archive database (such as MongoDB) to facilitate access and processing.

[0015] The data backup and recovery unit adopts a data backup and recovery mechanism to ensure that data can be quickly restored when a failure occurs.

[0016] Furthermore, the data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL), and performs data processing on the upload and storage module based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, including file preprocessing and information extraction, file classification and matching, and compliance review. The file processing and information extraction: the system uses natural language processing technology to perform text parsing and information extraction, extracts key information from contracts and agreements and standardizes them. The specific formula of the text extraction process is as follows:

[0017] Extracted(D i (t))_Information=NLP_Model(Raw_Text)

[0018] The natural language processing (NLP) model annotates and extracts the original text based on the context and outputs standardized key information; Raw_Text represents the original unprocessed text; NLP_Model represents the applied natural language processing model; Extracted (D i (t))_Information represents the key information extracted from the original text.

[0019] Furthermore, the file classification and matching module includes a convolutional neural network (CNN) or a long short-term memory network (LSTM). The file classification and matching module classifies files through a convolutional neural network (CNN) or a long short-term memory network (LSTM). By training the classification model, the system can automatically identify the file type and match it with other related files. The system also performs data matching through a machine learning algorithm to aggregate related files to ensure that all files related to the same project can be quickly associated together, and to ensure that all files related to the same project can be quickly associated together. In addition, for different application scenarios, corresponding subject directories are designed, which can be "generated with one click" when needed, and a set of files can be directly output for submission without repeated sorting;

[0020] The formula used by the archive classification and matching module is specifically analyzed as follows:

[0021] Document_category=CNN / LSTM_Classifier(Document_Text)

[0022] Document_Text is the text content of the file input; CNN / LSTM_Classifier represents the classifier trained using convolutional neural network (CNN) or long short-term memory network (LSTM); Document_category is the result of the system classifying the file according to the classification model.

[0023] Furthermore, the compliance review module: the system will use known regulatory texts to conduct automated compliance reviews on project files, and check whether each field in the file complies with the corresponding legal, financial, tax and other requirements through the NLP model and rule engine unit. The results of the compliance review will be fed back to the user. If non-compliant content is found, the system will automatically generate a report; the compliance review model is as follows:

[0024]

[0025] Where C(D) represents the overall compliance judgment result of file D; F i The i-th item in the file; R i Indicates the regulatory requirements related to the i-th clause, and Compliant indicates the i-th clause F i Comply with regulatory requirements? i , if it meets the requirements, return "Ture", otherwise return "False"; This means that all clauses in the file are checked one by one until all clauses pass the compliance judgment, and the file is considered compliant.

[0026] Furthermore, during the automated processing of archives in the evidence chain generation module, the system uses deep learning (DL) and natural language processing (NLP) technologies to automatically generate evidence chains. Each key document or clause automatically generates an evidence chain associated with the file through text extraction, information comparison and compliance verification. The evidence chain includes the review history of the archive, compliance verification results and possible legal basis, ensuring the traceability and compliance of the entire project archive.

[0027] Furthermore, the risk assessment and early warning module has a built-in risk assessment model, which analyzes key information in project files, combines historical data and pattern recognition technology, and evaluates the potential risks of the project in real time. The risk assessment uses regression analysis or decision tree algorithm to calculate the project risk score:

[0028]

[0029] in represents the risk score, represents the weight of the risk factor, Represents the characteristics of each risk factor; set the risk score R 阈值 , when the risk score is higher than the threshold, the system will issue an early warning signal: R risk >R 阈值 .

[0030] Furthermore, the workflow and task scheduling module: the system intelligently schedules various tasks in project management through a rule-based task scheduling engine, which optimizes the project process by dynamically adjusting the task sequence, priority and resource allocation. The workflow and task scheduling module includes a state machine model, which can be expressed as:

[0031] τ workflow ={τ1,τ2,...τ n}

[0032] where τ i Represents the i-th task node.

[0033] Furthermore, the user interaction and decision support module: the user interface (UI) of the system supports multi-terminal access, and users can access project archives through PC or mobile terminals to view archive processing results, compliance reports, and risk assessment reports;

[0034] The user interaction and decision support module interface is simple and intuitive, supporting search, filtering and quick navigation to help users quickly find key information. In addition, the system also provides decision support tools to help users make more accurate decisions based on risk warnings, compliance reports and other information.

[0035] Furthermore, according to the method for integrating public opinion data of financial activities based on knowledge graph according to claim 9, it is characterized in that the intelligent management platform for investment and financing project archive evidence based on big data according to any one of claims 1 to 9 is adopted, and the steps are as follows:

[0036] Step 1: First, the system uploads and preprocesses the archives through the upload and storage module, converting project files of different formats into structured data;

[0037] Step 2: The data processing and intelligent analysis module uses natural language processing (NLP) and machine learning (ML) algorithms to automatically extract key information and perform file classification, data matching and compliance review based on the uploaded and stored module data;

[0038] Step 3: The system then conducts automated review based on the compliance model, generates a chain of evidence that meets legal requirements, and conducts risk assessment in real time. When potential risks are identified, the system triggers an early warning and provides recommendations.

[0039] Step 4: Combined with risk warning and legal review data, the system workflow and task scheduling module dynamically schedules workflows based on task priorities and dependencies to ensure efficient and orderly project management;

[0040] Step 5: The system provides decision support through the user interface to help managers make accurate decisions. The overall process is seamless, greatly improving efficiency, accuracy and compliance.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. In the intelligent management platform for investment and financing project archive evidence based on big data, the present invention automatically generates an evidence chain that meets compliance requirements through intelligent technology, saving manual time and improving accuracy;

[0043] 2. In the intelligent management platform for investment and financing project archive evidence based on big data, the present invention uses the automated compliance check function to ensure that project documents meet legal requirements in real time and reduce legal risks. OK, continue to supplement and improve the patent content, and then supplement from the aspects of the overall advantages of the system, application scenarios and technical implementation details;

[0044] 2. In the intelligent management platform for investment and financing project archive evidence based on big data, the system can efficiently perform document classification and data matching based on machine learning and data matching algorithms, helping users to quickly find relevant information and avoiding the inefficiency and error rate of traditional manual search;

[0045] 3. In the intelligent management platform for investment and financing project archive evidence based on big data, through automated workflow scheduling, the system can flexibly adjust the execution order according to the priorities and dependencies of different tasks, ensure the smoothness of the project management process, avoid human interference, and improve the efficiency and quality of task completion;

[0046] 4. In the intelligent management platform for investment and financing project archive evidence based on big data of the present invention, the system can monitor potential risk factors in real time during the processing of project archives, predict project risks using machine learning and data mining technology, and issue warnings in time when risks exceed preset thresholds, ensuring that managers can take measures at the first time to reduce legal, financial and other risks;

[0047] 5. In the intelligent management platform for archival evidence of investment and financing projects based on big data of the present invention, by using natural language processing and artificial intelligence technology, the system can conduct intelligent review of documents such as contracts and agreements involved in investment and financing projects to ensure that all documents comply with the latest laws and regulations and avoid compliance omissions and legal proceedings. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0049] Figure 1 This is a schematic diagram of the overall steps of the intelligent management platform for investment and financing project archive evidence based on big data of the present invention;

[0050] Figure 2 This is a diagram of the evidence chain structure of the investment and financing project archive evidence intelligent management platform based on big data of the present invention; DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] like Figure 1 - Figure 2 As shown in the figure, the investment and financing project archive evidence intelligent management platform based on big data includes upload and storage module, data processing and intelligent analysis module, evidence chain generation module, risk assessment and early warning module, workflow and task scheduling module and user interaction and decision support module;

[0053] The upload and storage module includes a data upload unit, a data storage unit, and data backup and recovery. The data upload unit collects project files from different sources. After the files are uploaded to the system, they will first be pre-processed and then enter the data storage unit. The data storage unit uses a distributed database to store structured and unstructured data to ensure efficient access to massive data. At the same time, the data backup and recovery mechanism is used to ensure that data can be quickly restored when a failure occurs.

[0054] The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). The upload and storage module processes data based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, including file preprocessing and information extraction, file classification and matching, and compliance review. The system automatically processes and analyzes the project files of the data collected and stored by the upload and storage module through a variety of technical means such as file preprocessing and information extraction, file classification and matching, and compliance review, thereby improving work efficiency and accuracy, and ensuring that the files comply with relevant laws and regulations.

[0055] Evidence chain generation module: Based on the data processing and intelligent analysis module, deep learning (DL) and natural language processing (NLP) technologies are used to automatically generate evidence chains. Each key document or clause is automatically associated with the document through file preprocessing and information extraction, file classification and data matching, and compliance review.

[0056] Risk assessment and early warning module: Based on the evidence chain generation module and the data of historical projects, combined with the terms of the current contract, it generates a risk score for the project and determines whether there are legal or financial risks;

[0057] Workflow and task scheduling module: Based on the risk assessment and early warning module, the module shares the assessment and current workflow and tasks, and uses a rule-based task scheduling engine to automatically and dynamically adjust the order and priority of task execution based on the project archive content and progress;

[0058] User interaction and decision support module: users can access project archives through PC or mobile terminals, view archive processing results, compliance reports, risk assessment reports, and view system-recommended investment strategies or adjustment plans through decision support tools.

[0059] The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). Based on natural language processing (NLP), machine learning (ML) and deep learning (DL), the data processing and intelligent analysis module processes the data of the upload and storage modules through the following steps. The data upload collects different data PDF files, scanned archives or email multi-source files;

[0060] The data storage unit includes a distributed database to store structured and unstructured data. Distributed databases (such as Hadoop HDFS, Cassandra) store massive data to ensure data reliability and query speed. Unstructured data (such as scanned contract files, image files, etc.) are stored in archive databases (such as MongoDB) for easy access and processing.

[0061] The data backup and recovery unit adopts a data backup and recovery mechanism to ensure that data can be quickly restored when a failure occurs.

[0062] The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). The upload and storage module is processed based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, including file preprocessing and information extraction, file classification and matching, and compliance review. File processing and information extraction: The system uses natural language processing technology to perform text parsing and information extraction, extract key information from contracts and agreements, and standardize them. The specific formula of the text extraction process is as follows:

[0063] Extracted(D i (t))_Information=NLP_Model(Raw_Text)

[0064] The natural language processing (NLP) model annotates and extracts the original text based on the context and outputs standardized key information; Raw_Text represents the original unprocessed text; NLP_Model represents the applied natural language processing model; Extracted (D i (t))_Information represents the key information extracted from the original text.

[0065] Through this model, the original text is converted into standardized structured information to facilitate subsequent analysis and processing. At the same time, the application scenarios of archival evidence are determined, including various audit applications, various performance appraisal applications, and various settlement applications. In the later stage, other application scenarios such as other supervision and inspection, engineering claims, etc. can be added as needed. There is currently no authoritative research on the specific application scenarios. Therefore, this is a systematic summary and application.

[0066] The archive classification and matching module includes a convolutional neural network (CNN) or a long short-term memory network (LSTM). The archive classification and matching module classifies archives through a convolutional neural network (CNN) or a long short-term memory network (LSTM). By training the classification model, the system can automatically identify the archive type and match it with other related archives. The system also matches data through a machine learning algorithm and aggregates related archives to ensure that all files related to the same project can be quickly associated together. In addition, for different application scenarios, corresponding subject directories are designed, which can be "generated with one click" when needed, and a set of archives can be directly output for submission without repeated sorting;

[0067] The formula used by the archive classification and matching module is analyzed in detail as follows:

[0068] Document_category=CNN / LSTM_Classifier(Document_Text)

[0069] Document_Text is the text content of the file input; CNN / LSTM_Classifier represents the classifier trained using convolutional neural network (CNN) or long short-term memory network (LSTM); Document_category is the result of the system classifying the file according to the classification model.

[0070] The archive classification and matching module automatically identifies the type of archives through trained classifiers, and can effectively classify different types of archives and match them with other related archives. The archives are divided into audit scenarios (including internal audit, social audit, government audit, etc.), performance appraisal scenarios (construction performance appraisal, operation performance appraisal, mid-term evaluation, post-project evaluation, etc.), settlement scenarios (including process settlement, completion settlement, financial settlement, etc.), and other scenarios (such as superior supervision and inspection, engineering claims, dispute resolution, litigation, etc.). For different application scenarios, corresponding subject directories are designed, which can be "generated with one click" when needed, and a set of archives can be directly output for submission without repeated sorting. Taking government audit as an example, it usually includes basic construction procedure audit, engineering management audit, performance evaluation audit, etc. According to the application scenario, it should be able to automatically form a chain of evidence so that the program can run automatically to form a "one-click generation" effect. Its basic paradigm logic is as follows Figure 2 .

[0071] Compliance review module: The system will use known regulatory texts to conduct automated compliance reviews on project files. Through the NLP model and rule engine unit, it will check whether each field in the file complies with the corresponding legal, financial, tax and other requirements. The results of the compliance review will be fed back to the user. If non-compliant content is found, the system will automatically generate a report. The compliance review model is as follows:

[0072]

[0073] Where C(D) represents the overall compliance judgment result of file D; F i The i-th item in the file; R i Indicates the regulatory requirements related to the i-th clause, and Compliant indicates the i-th clause F i Comply with regulatory requirements? i , if it meets the requirements, return "Ture", otherwise return "False"; This means that all clauses in the file are checked one by one until all clauses pass the compliance judgment, and the file is considered compliant.

[0074] The compliance review model reviews each clause one by one and combines the compliance results of all clauses to determine the compliance of the entire file. If all clauses in the file comply with the regulations, it returns "compliant"; if any clause does not comply with the regulations, it returns "non-compliant";

[0075] During the automated processing of archives in the evidence chain generation module, the system uses deep learning (DL) and natural language processing (NLP) technologies to automatically generate evidence chains. Each key document or clause automatically generates an evidence chain associated with the file through text extraction, information comparison and compliance verification. The evidence chain includes the review history of the archive, compliance verification results and possible legal basis, ensuring the traceability and compliance of the entire project archive.

[0076] When the terms in the contract are extracted and compliance checked, the system will generate a corresponding chain of evidence to record all review processes, including compliance checks, review report generation, signing status, etc. The chain of evidence will be automatically saved and provide a legal basis for the archive to ensure that complete proof can be provided in the future. The data of each application scenario is automatically matched, and a clear and corresponding complete chain of evidence is designed in advance. The risk assessment and early warning module has a built-in risk assessment model. By analyzing the key information in the project documents, combined with historical data and pattern recognition technology, the potential risks of the project are assessed in real time. Risk assessment uses regression analysis or decision tree algorithm to calculate the project risk score:

[0077]

[0078] in represents the risk score, represents the weight of the risk factor,

[0079] Represents the characteristics of each risk factor; set the risk score R 阈值 , when the risk score is higher than the threshold, the system will issue an early warning signal: R risk >R 阈值 .

[0080] The risk assessment and early warning module uses data mining and pattern recognition technology to monitor and evaluate the risks in investment and financing projects in real time. Based on the data in the project files (such as investment amount, project progress, signing date, etc.), the system predicts potential risks through regression models, decision tree algorithms, etc. When the risk score exceeds the preset threshold, the system will automatically issue a warning and recommend preventive measures.

[0081] Workflow and task scheduling module: The system intelligently schedules various tasks in project management through a rule-based task scheduling engine. The engine optimizes the project process by dynamically adjusting the task sequence, priority and resource allocation. The workflow and task scheduling module includes a state machine model, which can be expressed as:

[0082] τ workflow ={τ1,τ2,...τ n}

[0083] where τ i Represents the i-th task node.

[0084] The workflow and task scheduling module adopts a rule-based task scheduling engine, which automatically adjusts the task execution order and priority according to the project file content and progress. The workflow model is managed through a graph structure (such as state machine, flow chart, etc.) to ensure that various tasks (such as file review, data extraction, risk assessment) are completed efficiently and orderly. This technology is more intelligent and automated.

[0085] User interaction and decision support module: The system's user interface (UI) supports multi-terminal access. Users can access project archives through PC or mobile terminals to view archive processing results, compliance reports, and risk assessment reports;

[0086] The user interaction and decision support module interface is simple and intuitive, supporting search, filtering and quick navigation to help users quickly find key information. In addition, the system also provides decision support tools to help users make more accurate decisions based on risk warnings, compliance reports and other information.

[0087] The method for integrating public opinion data of financial activities based on knowledge graph is characterized in that the intelligent management platform for investment and financing project archive evidence based on big data according to any one of claims 1 to 9 is adopted, and the steps are as follows:

[0088] Step 1: First, the system uploads and preprocesses the archives through the upload and storage module, converting project files of different formats into structured data;

[0089] Step 2: The data processing and intelligent analysis module uses natural language processing (NLP) and machine learning (ML) algorithms to automatically extract key information and perform file classification, data matching and compliance review based on the uploaded and stored module data;

[0090] Step 3: The system then conducts automated review based on the compliance model, generates a chain of evidence that meets legal requirements, and conducts risk assessment in real time. When potential risks are identified, the system triggers an early warning and provides recommendations.

[0091] Step 4: Combined with risk warning and legal review data, the system workflow and task scheduling module dynamically schedules workflows based on task priorities and dependencies to ensure efficient and orderly project management;

[0092] Step 5: The system provides decision support through the user interface to help managers make accurate decisions. The overall process is seamless, greatly improving efficiency, accuracy and compliance.

[0093] Working principle: Based on artificial intelligence, big data analysis and automation technology, it aims to optimize the management process of investment and financing project archives. First, the system converts project files of different formats (such as contracts, financial reports) into structured data through archive upload and preprocessing modules. Then, it uses natural language processing (NLP) and machine learning (ML) algorithms to automatically extract key information and perform archive classification, data matching and compliance review. Next, the system performs automated review based on the compliance model, generates a chain of evidence that meets legal requirements, and conducts risk assessment in real time. When potential risks are identified, the system triggers an early warning and provides suggestions. In addition, through automated workflow management, the system dynamically schedules workflows based on task priorities and dependencies. process to ensure efficient and orderly project management. Finally, the system provides decision support through the user interface to help managers make accurate decisions. The overall process is seamlessly connected, which greatly improves efficiency, accuracy and compliance. The intelligent investment and financing project archive management system of the present invention fully integrates artificial intelligence, big data, machine learning and natural language processing technology, and automatically completes the generation, classification, compliance review, evidence chain generation, risk assessment and other tasks of project archives. Through this system, enterprises can effectively improve the efficiency of archive management (various audit applications, various performance appraisal applications, various settlement applications), while reducing legal risks and improving the accuracy of investment decisions, providing an intelligent and automated solution for the investment and financing industry.

[0094] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent management platform for investment and financing project archive evidence based on big data, characterized by: It includes upload and storage module, data processing and intelligent analysis module, evidence chain generation module, risk assessment and early warning module, workflow and task scheduling module and user interaction and decision support module; The upload and storage module includes a data upload unit, a data storage unit and data backup and recovery. The data upload unit collects project files from different sources. After the files are uploaded to the system, they will first be pre-processed and then enter the data storage unit. The data storage unit uses a distributed database to store structured and unstructured data to ensure efficient access to massive data. At the same time, the data backup and recovery mechanism is used to ensure that data can be quickly restored when a failure occurs; The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). The upload and storage module is processed based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data uploaded to the pre-storage module through the following steps, including file pre-processing and information extraction, file classification and matching and compliance review. The system automatically processes and analyzes the project files of the data collected and stored by the upload and storage module through multiple technical means such as file pre-processing and information extraction, file classification and matching and compliance review, thereby improving work efficiency and accuracy, and ensuring that the files comply with relevant laws and regulations. The evidence chain generation module: based on the data processing and intelligent analysis module, uses deep learning (DL) and natural language processing (NLP) technology to automatically generate an evidence chain. Each key document or clause is automatically generated with an evidence chain associated with the document through file preprocessing and information extraction, file classification and data matching, and compliance review. The risk assessment and early warning module: based on the evidence chain generation module and the data of historical projects, combined with the terms of the current contract, generates a risk score for the project and determines whether there are legal or financial risks; The workflow and task scheduling module: based on the risk assessment and early warning module’s shared assessment and current workflow and tasks, adopts a rule-based task scheduling engine to automatically and dynamically adjust the task execution order and priority according to the project file content and progress; In the user interaction and decision support module, users can access project archives through PC or mobile terminals, view archive processing results, compliance reports, risk assessment reports, and view system-recommended investment strategies or adjustment plans through decision support tools.

2. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 1 is characterized in that: The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL), and performs data processing on the upload and storage module based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps. The data upload collects different data PDF files, scanned archives or email multi-source files; The data storage unit includes a distributed database to store structured and unstructured data. The distributed database (such as Hadoop HDFS, Cassandra) stores massive data to ensure data reliability and query speed. The unstructured data (such as scanned contract files, picture files, etc.) is stored in an archive database (such as MongoDB) to facilitate access and processing. The data backup and recovery unit adopts a data backup and recovery mechanism to ensure that data can be quickly restored when a failure occurs.

3. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 2 is characterized in that: The data processing and intelligent analysis module includes natural language processing (NLP), machine learning (ML) and deep learning (DL). The upload and storage module is processed based on natural language processing (NLP), machine learning (ML) and deep learning (DL). The data processing and intelligent analysis module processes the data of the upload and storage module through the following steps, including file preprocessing and information extraction, file classification and matching, and compliance review. The file processing and information extraction: the system uses natural language processing technology to perform text parsing and information extraction, extracts key information from contracts and agreements and standardizes them. The specific formula of the text extraction process is as follows: Extracted(D i (t))_Information=NLP_Model(Raw_Text) The natural language processing (NLP) model annotates and extracts the original text based on the context and outputs standardized key information; Raw_Text represents the original unprocessed text; NLP_Model represents the applied natural language processing model; Extracted (D i (t))_Information represents the key information extracted from the original text.

4. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 3 is characterized in that: The file classification and matching module includes a convolutional neural network (CNN) or a long short-term memory network (LSTM). The file classification and matching module classifies files through a convolutional neural network (CNN) or a long short-term memory network (LSTM). By training the classification model, the system can automatically identify the file type and match it with other related files. The system also matches data through a machine learning algorithm and aggregates related files to ensure that all files related to the same project can be quickly associated together. In addition, corresponding subject directories are designed for different application scenarios, which can be "generated with one click" when needed, and a set of files can be directly output for submission without repeated sorting; The formula used by the archive classification and matching module is specifically analyzed as follows: Document_category=CNN / LSTM_Classifier(Document_Text) Document_Text is the text content of the file input; CNN / LSTM_Classifier represents the classifier trained using convolutional neural network (CNN) or long short-term memory network (LSTM); Document_category is the result of the system classifying the file according to the classification model.

5. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 3 is characterized in that: The compliance review module: The system will use known regulatory texts to conduct automated compliance reviews on project files. Through the NLP model and rule engine unit, it will check whether each field in the file complies with the corresponding legal, financial, tax and other requirements. The results of the compliance review will be fed back to the user. If non-compliant content is found, the system will automatically generate a report. The compliance review model is as follows: Where C(D) represents the overall compliance judgment result of file D; F i The i-th item in the file; R i Indicates the regulatory requirements related to the i-th clause, and Compliant indicates the i-th clause F i Comply with regulatory requirements? i , if it meets the requirements, return "Ture", otherwise return "False"; This means that all clauses in the file are checked one by one until all clauses pass the compliance judgment, and the file is considered compliant.

6. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 5 is characterized in that: During the automated processing of archives in the evidence chain generation module, the system uses deep learning (DL) and natural language processing (NLP) technologies to automatically generate evidence chains. Each key document or clause is automatically generated through text extraction, information comparison and compliance verification. The evidence chain associated with the document includes the review history of the archive, compliance verification results and possible legal basis, ensuring the traceability and compliance of the entire project archive.

7. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 1 is characterized in that: The risk assessment and early warning module has a built-in risk assessment model. By analyzing key information in project files, combined with historical data and pattern recognition technology, it can assess the potential risks of the project in real time. The risk assessment uses regression analysis or decision tree algorithm to calculate the project risk score: in represents the risk score, represents the weight of the risk factor, Represent the characteristics of each risk factor; set the risk score R 阈值 , when the risk score is higher than the threshold, the system will issue an early warning signal: R risk >R 阈值 .

8. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 7 is characterized in that: The workflow and task scheduling module: The system intelligently schedules various tasks in project management through a rule-based task scheduling engine. The engine optimizes the project process by dynamically adjusting the task sequence, priority and resource allocation. The workflow and task scheduling module includes a state machine model, which can be expressed as: t workflow ={τ1,τ2,......τ n } where τ i Represents the i-th task node.

9. The intelligent management platform for investment and financing project archive evidence based on big data according to claim 1 is characterized in that: The user interaction and decision support module: The system's user interface (UI) supports multi-terminal access. Users can access project archives through PC or mobile terminals to view archive processing results, compliance reports, and risk assessment reports; The user interaction and decision support module interface is simple and intuitive, supporting search, filtering and quick navigation to help users quickly find key information. In addition, the system also provides decision support tools to help users make more accurate decisions based on risk warnings, compliance reports and other information.

10. The method for integrating public opinion data of financial activities based on knowledge graph according to claim 9 is characterized in that: Using the investment and financing project archive evidence intelligent management platform based on big data as described in any one of claims 1 to 9 above, the steps are as follows: Step 1: First, the system uploads and preprocesses the archives through the upload and storage module, converting project files of different formats into structured data; Step 2: The data processing and intelligent analysis module uses natural language processing (NLP) and machine learning (ML) algorithms to automatically extract key information and perform file classification, data matching and compliance review based on the uploaded and stored module data; Step 3: The system then conducts automated review based on the compliance model, generates a chain of evidence that meets legal requirements, and conducts risk assessment in real time. When potential risks are identified, the system triggers an early warning and provides suggestions. Step 4: Combining risk warning and legal review data, the system workflow and task scheduling module dynamically schedules workflows based on task priorities and dependencies to ensure efficient and orderly project management; Step 5: The system provides decision support through the user interface to help managers make accurate decisions. The overall process is seamless, greatly improving efficiency, accuracy and compliance.

Images