Banking system research and development project management system

Abstract

The invention discloses a bank system research and development project management system, and the system comprises an intelligent management platform which generates a compliance check list and a difference report through the real-time butt joint of a supervision database; the risk prediction module is used for predicting risk levels such as research and development cycle delay and resource hyperbranched, triggering early warning and suggesting a deployment scheme; the block chain evidence storage and tracing module is used for storing the hash value and the operation record of the key node document in a chaining manner, and solidifying a timestamp and a person in charge; the task collaboration module is used for analyzing a demand document by using NLP, automatically disassembling the demand document into an executable task tree, analyzing manpower and equipment conflicts through a resource occupation graph, and intelligently generating a deployment scheme; and the security control module is used for implementing hierarchical authority based on role and data sensitivity, and sensitive operation needs dual verification of biometric authentication and dynamic token to prevent unauthorized access. The system is high in risk resistance, high in overall research and development efficiency and fast in data management and auditing, and can meet the requirements in the research and development process of a bank system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bank system development, and particularly relates to a bank system research and development project management system. BACKGROUND

[0002] In the wave of financial informatization, bank system research and development project management has undergone a long evolution. In the early stage, bank system development relies on manual recording and manual experience, and the management of project progress, resource allocation and other aspects is scattered and disordered. Document management is also in disorder, and compliance check basically relies on manual comparison of policy documents, which is low in efficiency and easy to make mistakes. With the expansion of bank business and the preliminary penetration of information technology, basic project management tools such as simple progress tracking software and resource management table are introduced, which completes the transformation from disorder to preliminary standardization, but these tools are single in function and poor in integration, and it is difficult to meet the increasing complexity of bank system research and development.

[0003] Currently, a variety of mainstream technologies are widely used in the field of bank system research and development project management. On the one hand, mature project management software such as Jira and Trello provides visual board, agile development process support, task allocation and other functions to help team efficient collaboration; on the other hand, agile development methodology is popular, which emphasizes iterative development and rapid response to demand changes, so that the bank system can keep pace with the market; at the same time, data warehouse and business intelligence tools are combined to preliminarily realize the collection and simple analysis of project data, providing certain basis for decision-making. These technologies indeed play an important role in improving project management efficiency and promoting team collaboration, making the promotion of bank system research and development project more orderly and transparent.

[0004] However, the existing mainstream technologies have exposed many problems in the deep application in the financial field. Firstly, the compliance guarantee ability is limited, and it is difficult to cope with the increasingly complex regulatory policies by manual screening, which easily leads to compliance risks; secondly, risk early warning is lagging, and the existing tools are mostly passive in response after the problem appears, and cannot accurately predict the hidden troubles such as research and development cycle delay and resource overspending in advance; thirdly, the data island phenomenon is serious, and the project management system, human resource system and other systems are independent, and the data is difficult to share and integrate, causing information fragmentation and affecting the comprehensiveness of decision-making; fourthly, the resource allocation is not intelligent, and is only based on simple rules, without fully considering dynamic factors such as personnel skills and equipment status, resulting in frequent waste of resources and conflicts. These problems restrict the further improvement of bank system research and development project management level, and it is difficult to meet the strict requirements of modern banks for high efficiency, low risk and strong compliance. SUMMARY

[0005] The present application provides a bank system research and development project management system to solve the technical problems of poor risk resistance of existing management system, low overall research and development efficiency, slow data management and audit.

[0006] To solve the above technical problems, the present application provides the following technical solutions: The present application provides a bank system R&D project management system, comprising: An intelligent management platform generates a compliance check list and a difference report by real-time interfacing with a regulatory database, scanning sensitive words in project documents using NLP technology, marking high-risk tasks and recommending rectification schemes; A risk prediction module predicts the risk level of R&D cycle delays, resource overruns, etc. based on machine learning analysis of historical project data, triggers an early warning and suggests a deployment plan; A blockchain storage and traceability module stores the hash value and operation record of key node documents on the chain, solidifies the timestamp and responsible person, and supports regulatory agencies to quickly access the full-process audit log through a private key; A task coordination module uses NLP to analyze requirement documents, automatically disassembles them into executable task trees, analyzes human and equipment conflicts through resource occupation graph analysis, and intelligently generates a deployment plan; A security control module implements hierarchical permissions based on roles and data sensitivity, and requires biometric authentication and dynamic token double verification for sensitive operations to prevent unauthorized access.

[0007] Further, an implementation process of a bank system R&D project management system, comprising the following steps, Initialize system parameters, configure the functions and permissions of each module, including interfacing with regulatory database account information, setting basic permissions for each role, and initializing the blockchain node; Real-time interface the regulatory database through the intelligent management platform, use the preset data collection rules and NLP algorithm to analyze and scan the sensitive words of the obtained policy update information, generate a compliance check list and a difference report based on the rule engine, and send a to-do list notification to the relevant project manager through the system's built-in message push service; Use the risk prediction module to evaluate the project, the model loads historical project data for self-training and optimization, and outputs the risk level prediction result in real time according to the current project progress, when the risk reaches the warning threshold, triggers the warning process, and provides a deployment suggestion based on the current resource state; Store key node documents through the blockchain storage and traceability module, automatically calculate the document hash value and store related information on the chain, and record operation logs to ensure that each key operation has a traceable record; use the task coordination module to analyze, disassemble and recommend paths for requirement documents, project managers confirm the recommended task tree and development path in the system, tasks are assigned to specific executors, and the resource analysis unit monitors resource occupation in real time, and if a conflict occurs, it will automatically adjust or prompt manual intervention; The security control module implements security control throughout the whole process, biological recognition and dynamic token verification are carried out when personnel log in the system, access to sensitive resources is authenticated again, and all operation requests are released after security policy inspection, so that the system security is guaranteed. The visual display unit obtains data from each business module in real time, renders and displays according to preset visual rules, and assists managers in decision-making. According to the feedback and data analysis results of the system, the project team continuously optimizes the execution process of the project until the completion of the project, so as to realize efficient management and smooth delivery of the bank system development project.

[0008] The technical scheme provided by the application has at least the following beneficial effects: The following are 3 beneficial effects according to the content of the claim: The intelligent management platform of the application can real-time interface with a supervision database, automatically capture and analyze policy updates, scan sensitive words in project documents by combining NLP technology, timely discover compliance risk points in the project, generate compliance inspection lists and difference reports, mark high-risk tasks and recommend rectification schemes, ensure that the bank system development project meets regulatory requirements, avoid huge fines, business restrictions and other risks due to violations, protect the stable operation of the bank, and the risk prediction module can predict the risk level of development cycle delay, resource overruns and other risks in advance based on machine learning analysis of historical project data, real-time monitor resource occupation conflicts and trigger early warnings, provide sufficient time for project management to take measures to mitigate risks, reduce the possibility of project failure, and enhance the risk resistance of the bank in the system development process.

[0009] The task coordination module of the application can accurately convert requirements into specific development tasks by using NLP to analyze requirement documents, automatically decompose them into executable task trees, and recommend development paths based on bank business rules, reduce errors and omissions in manual task decomposition, recommend optimal development paths to avoid detours in the development process, shorten the project development cycle, analyze human and equipment conflicts through resource occupation graph, intelligently generate deployment schemes, realize fine management and reasonable allocation of development resources, improve resource utilization, avoid resource idling and waste, enable the bank to complete more development tasks with limited resources, and improve overall development efficiency.

[0010] In the task coordination module of the application, the resource analysis unit can evaluate the resource matching degree of outsourcing tasks, analyze the skill level, available time window and cost-effectiveness of outsourcing teams, and ensure that the resource demand of outsourcing tasks is met, which can effectively reduce the uncertainty and potential risks in the outsourcing process, ensure the smooth execution of outsourcing tasks, and improve the overall controllability of the project.

[0011] The security control module of the application implements hierarchical permissions based on roles and data sensitivity, and sensitive operations need to be verified by biological authentication and dynamic token, which effectively prevents unauthorized access, protects the confidentiality and integrity of core data assets such as customer information and transaction data in the bank system development process, reduces the risk of data leakage, maintains the reputation and customer trust of the bank, the blockchain storage and traceability module stores the hash value and operation record of the key node document on the chain, solidifies the timestamp and the person in charge, supports the regulatory agencies to quickly review the whole process audit log through the private key, ensures the transparency and traceability of the development process, facilitates the quick positioning of responsibility when problems occur, and also provides convenience for the bank's internal audit work, improves the audit efficiency, and enhances the compliance and credibility of the bank in data management and audit. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0013] Figure 1 The bank system research and development project management system system block diagram provided by the embodiment of the application. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical scheme and advantages of the application more clear, the embodiments of the application will be further described in detail below with reference to the drawings.

[0015] Please refer to Figure 1 The bank system research and development project management system system block diagram provided by the embodiment of the application. The embodiment provides a bank system research and development project management system, which comprises: 1. Intelligent management platform By real-time docking with the regulatory database, combining NLP technology to scan sensitive words in project documents, generating compliance check list and difference report, marking high-risk tasks and recommending rectification scheme.

[0016] It includes data collection unit, policy analysis unit, sensitive word scanning unit and report generation unit. The data collection unit adopts a distributed crawler architecture based on the open API standard protocol of the regulatory database, uses multi-threading technology to ensure that the incremental data is captured within the first time window of regulatory data update, and stores the data in the message queue Kakfa to prevent data loss due to sudden increase in data volume and ensure data collection continuity. The policy analysis unit uses a deep learning language model based on the Transformer architecture to pre-train the complex structure and professional terms of financial policy text, and then fine-tune it according to the business characteristics of Hu Zhou Bank. The model can accurately identify the key elements in the policy and convert the policy provisions into structured data to provide accurate basis for subsequent compliance checks. The sensitive word scanning unit constructs a sensitive word knowledge graph based on the bank business scenario, stores sensitive words and their associated relationships in the graph database Neo4j, and uses graph neural network technology to dynamically evaluate the risk level of sensitive words according to the context semantics of the document, avoiding false positives caused by simple keyword matching. The report generation unit compares the compliance requirements, sensitive word risk assessment results and project documents based on the rule engine Drools, uses pre-set rule templates to automatically generate a compliance check list containing difference details and risk distribution heat map, and uses a deep learning text generation model to customize a rectification plan for each high-risk task to assist the project team in efficiently responding to policy changes.

[0017] 2. Risk prediction module Based on machine learning analysis of historical project data, predict the risk level of R&D cycle delay and resource overspending, trigger early warning and suggest allocation solutions.

[0018] It includes data storage unit, machine learning analysis unit, risk assessment unit and early warning unit. The data storage unit adopts distributed columnar storage system HBase, and classifies and stores the structured and unstructured parts in the historical project data. For unstructured data such as requirement documents and meeting minutes, the text vectorization technology Word2Vec is used to convert them into analyzable feature vectors, which are stored uniformly with structured data for subsequent model training, fast retrieval and batch reading. The machine learning analysis unit uses long short-term memory network to process project time series data, captures the fluctuation law of R&D cycle, and combines gradient boosting tree GBDT algorithm to fit multi-dimensional features of resource consumption, builds a hybrid model, and continuously optimizes model hyperparameters through cross-validation to improve prediction accuracy. The risk assessment unit builds a risk assessment model based on Bayesian network, integrates the risk factors and their correlation of R&D cycle delay and resource overrun into the probabilistic graph model, dynamically updates the prior probability according to the current project status, and calculates the risk level posterior probability in real time to provide quantitative basis for early warning. The early warning unit uses real-time stream computing framework Flink to monitor project resource occupation indicators in real time, sets multi-level early warning thresholds, and when the resource conflict degree reaches the threshold, sends early warning information to the relevant person in charge through the message push service RabbitMQ, and calls the resource allocation algorithm to provide project team with scheme suggestions including manpower allocation and equipment sharing based on the current resource occupation graph, to ensure the smooth progress of the project.

[0019] 3. Blockchain storage and traceability module The hash value and operation record of the key node document are stored on the chain, and the time stamp and responsible person are fixed. Supervising agencies can quickly review the whole process audit log through private key.

[0020] It includes a hash calculation unit, a blockchain storage unit and an audit log management unit. The hash calculation unit adopts SHA-256 algorithm combined with Merkle tree structure to perform multi-layer hash operation on the key node document, generating a document fingerprint with high security, ensuring that any minor tampering can be discovered in time, and using GPU acceleration technology to improve the efficiency of hash calculation, meeting the performance requirements of high-frequency updates of bank project documents. The blockchain storage unit is based on the consortium chain architecture Hyperledger Fabric, and builds a multi-party participation evidence storage network including regulatory agencies, bank management and project teams. Each block records the hash value of the key node document, operation record, timestamp and the information of the digital signature of the responsible person. The legality of the transaction is automatically verified by the smart contract to ensure that the data on the chain is real and reliable and cannot be tampered with, providing a solid foundation for subsequent audits. The audit log management unit designs an encrypted index mechanism, encrypts the relevant indexes of the audit log using the private key of the regulatory agency, ensures that only the regulatory agency can quickly locate the required log information within the authorized range, and uses zero-knowledge proof technology to protect the sensitive information of the project while enabling the regulatory agency to verify the integrity and authenticity of the log, achieving efficient audit supervision.

[0021] 4. Task coordination module The NLP is used to analyze the requirement document, which is automatically disassembled into an executable task tree, and the resource occupation graph is used to analyze the conflict between manpower and equipment, and an intelligent deployment scheme is generated.

[0022] The system comprises a demand analysis unit, a task decomposition unit, a path recommendation unit and a resource analysis unit. The demand analysis unit uses an improved BERT model to conduct deep semantic analysis on demand documents, conducts special training on common business logic and terms in bank demand documents, accurately identifies functional points, non-functional constraints and associated business rules in the demand, and provides accurate semantic understanding for subsequent task decomposition. The task decomposition unit uses the hierarchical task network (HTN) planning algorithm to decompose the parsed demand into an executable task tree according to the hierarchical structure of the bank business process, each task node contains clear input and output parameters, preconditions and postconditions, and an estimated workload, and uses a dependency graph to clearly show the complex dependency between tasks, ensuring that the task execution order is reasonable. The path recommendation unit constructs a knowledge base containing multiple bank business architecture templates, combines the current project demand characteristics and the current bank technology stack, and uses a reinforcement learning algorithm to recommend the optimal development path for the task tree. The algorithm continuously optimizes the recommendation strategy based on historical project path execution feedback to improve development efficiency and quality. The resource analysis unit uses graph neural network technology to construct a resource occupation graph, treats human and equipment resources as graph nodes, and treats task resource demand relationships as graph edges. The graph convolutional neural network (GCN) analyzes resource conflict propagation paths, predicts potential conflict ranges, intelligently generates deployment schemes, and considers factors such as personnel skill matching and equipment availability time windows to achieve efficient resource utilization.

[0023] 5. Security control module Based on the role and data sensitivity, hierarchical permissions are implemented, and sensitive operations require dual verification of biometric authentication and dynamic token to prevent unauthorized access.

[0024] The system comprises a permission management unit, an authentication unit and an access control unit. The permission management unit extends the data sensitivity dimension based on the role-based access control (RBAC) model, constructs a multi-level permission system, sets access levels according to the sensitivity of different bank business systems and project documents, and combines dynamic permission adjustment strategies to update permission configurations in real time according to project phases and personnel position changes, ensuring accurate and adaptive permission allocation. The authentication unit uses multi-modal biometric recognition technology to integrate multiple biometric features such as fingerprints, faces and voiceprints, uses a deep learning feature extraction model to improve biometric recognition accuracy and anti-forgery capability, and combines time-synchronized one-time password (TOTP) dynamic token technology to generate dynamic verification codes with high security, ensuring the security of sensitive operations through dual verification mechanisms to prevent unauthorized operations by illegal personnel. The access control unit uses software-defined security (SDS) technology to construct a virtualized security boundary, dynamically adjusts network access strategies based on permission management and authentication results, encrypts sensitive data transmission channels, uses an intrusion detection system (IDS) to monitor access behavior in real time, and uses machine learning anomaly detection algorithms to detect and block abnormal access requests in a timely manner, providing comprehensive protection for system security.

[0025] It should be noted that information is exchanged between modules through a data transmission interface based on the message middleware RocketMQ and the API gateway. The data transmission interface adopts the RESTfulAPI specification to ensure the simplicity and universality of inter-module calls, and uses the API gateway for request routing, load balancing, and authentication management to ensure secure and reliable data interaction. To achieve data sharing and collaborative work, each module follows the unified data exchange format JSON during data transmission, standardizes the definition of data fields in combination with the data dictionary, and implements asynchronous communication through the message middleware to ensure the timeliness and integrity of data transmission in high-concurrency scenarios, avoid blocking caused by dependencies between modules, thereby ensuring the effective operation of the overall system functions and improving the efficiency and quality of bank system R&D project management.

[0026] It's important to note that the intelligent management platform integrates with other internal bank management systems, including the project management system (based on the agile development framework Jira) and the human resources management system (based on the SAP HANA platform). By building a unified data integration layer and utilizing the ETL tool Informatica PowerCenter for data extraction, transformation, and loading, data interoperability is achieved. For example, regarding project progress, the intelligent management platform obtains task status updates from the project management system. Combined with personnel skills and schedule data from the human resources management system, it utilizes the critical path method (CPM) from operations research to reassess overall project progress, promptly identifying potential delay risks and providing feedback to relevant systems, triggering appropriate adjustments and further enhancing the comprehensiveness and accuracy of the bank's system R&D project management.

[0027] It should be noted that the system also includes a visualization display unit, which utilizes a micro-frontend architecture and integrates visualization components for different business dimensions via the AntV data visualization component as plug-ins. For project progress visualization, a Gantt chart visually compares task plans with actual execution, and a traffic light color warning mechanism is used to highlight progress deviations. For risk visualization, a risk heat map is constructed, dynamically colored based on risk level and impact range, assisting managers in quickly locating key risk points. Resource utilization is displayed using a Sankey diagram, clearly illustrating resource flow and allocation. WebSocket real-time push technology ensures timely updating of displayed data, facilitating decision-making and monitoring by managers, allowing them to identify problems and take appropriate measures.

[0028] It should be noted that the implementation process is introduced below.

[0029] 1. Initialization preparation phase The system configuration team is responsible for initializing system parameters, including accurate configuration of interface regulatory database account information to ensure smooth connection to the regulatory database and access to the latest regulatory policy data. At the same time, according to the organizational structure and business needs of the bank system, set the basic permissions of each role, and clearly define the operation range and permission level of different roles in the system, such as project managers, developers, and testers, who have different data access and operation permissions. In addition, complete the initialization of the blockchain node, determine the initial architecture and participating nodes of the blockchain network, and lay the foundation for subsequent evidence storage and traceability functions. In this phase, the system's basic parameters also need to be optimized to adapt to the actual business scale and complexity of the bank system, ensuring that the system works efficiently and stably during operation.

[0030] 2. Data collection and compliance check phase The data collection unit of the intelligent management platform is started, and according to the preset data collection rules and frequency, it is connected to the regulatory database in real time. Taking advantage of multi-threading technology and distributed crawler architecture, it quickly captures incremental updates of regulatory data, and uses message queue Kafka for data caching to ensure the continuity and integrity of data collection, avoiding data loss due to large data volume or network fluctuations. The collected data will be transmitted to the policy analysis unit, which uses a deep learning language model based on the Transformer architecture to deeply analyze financial policy texts, accurately identify key elements, and store them as structured data. At the same time, the sensitive word scanning unit uses graph neural network technology combined with document context semantics based on the constructed bank business scenario sensitive word knowledge graph to comprehensively scan project documents and dynamically assess the risk level of sensitive word occurrence. The report generation unit uses rule engine Drools and deep learning text generation model to automatically generate detailed compliance check lists and difference reports based on compliance requirements, sensitive word risk assessment results, and project document content, which include detailed differences, risk distribution heat maps, and customized rectification plans for each high-risk task. The rectification plan details the rectification steps, responsible departments, and key information such as expected time consumption, providing clear rectification direction and operation guidelines for the project team to efficiently address compliance challenges brought by policy changes. The system's built-in message push service will send a to-do list notification to the relevant project manager after the report is generated, reminding them to pay attention to compliance issues and take appropriate rectification measures.

[0031] 3. Risk assessment and early warning phase The risk prediction module starts a comprehensive risk assessment of the project. The data storage unit stores historical project data in a classified manner, where unstructured data such as requirement documents and meeting minutes are converted into feature vectors through the text vectorization technology Word2Vec and stored in the distributed columnar storage system HBase together with structured data, so that the machine learning analysis unit can quickly retrieve and batch read these data for model training. The machine learning analysis unit uses a combination of long short-term memory networks and gradient boosting tree GBDT algorithms to build a hybrid model, which analyzes time series data and resource consumption multi-dimensional features of the project in depth, captures research and development cycle fluctuation patterns and resource consumption patterns. During model training, cross-validation is used to continuously optimize model hyperparameters and improve risk prediction accuracy. The risk assessment unit builds a risk assessment model based on Bayesian networks, integrates various risk factors and their associated relationships of research and development cycle delay and resource overspending into a probabilistic graph model, dynamically updates the prior probability based on the current project status, and calculates the posterior probability of the risk level in real time to provide quantitative basis for early warning. When the risk level reaches the early warning threshold, the early warning unit quickly starts, uses the real-time stream computing framework Flink to monitor project resource occupation indicators in real time, and once the resource conflict degree reaches the set multi-level early warning threshold, immediately sends early warning information to the relevant responsible person through the message push service RabbitMQ, reminding them to pay attention to project risks in a timely manner. At the same time, the early warning unit calls the resource allocation algorithm and provides suggestions for the project team based on the current resource occupation graph, including manpower allocation, equipment sharing and other schemes, to help them adjust resource allocation reasonably and ensure the smooth progress of the project, reducing the adverse effects of risks on the project.

[0032] 4. Evidence storage and traceability stage The blockchain storage and traceability module plays a role at the key nodes of the project. When the key node document is generated or updated, the hash calculation unit adopts the SHA-256 algorithm combined with the Merkle tree structure to perform multi-layer hash operation on the document to generate a document fingerprint with high security. Using GPU acceleration technology, it ensures that the hash calculation can be completed efficiently in the environment of high-frequency updates of bank project documents, and any minor tampering behavior can be detected in time. The blockchain storage unit records the calculated document hash value, operation record, timestamp, and information signed by the responsible person in the block based on the consortium chain architecture Hyperledger Fabric, and automatically verifies the legality of the transaction through the smart contract to ensure that the data on the chain is real and reliable and cannot be tampered with, providing a solid and reliable data foundation for subsequent audit work. The audit log management unit designs an encrypted index mechanism, encrypts the relevant index of the audit log using the private key of the supervisory agency, and ensures that only the supervisory agency can quickly and accurately locate the required log information within the authorized range. At the same time, with the help of zero-knowledge proof technology, the supervisory agency can effectively verify the integrity and authenticity of the log while protecting the sensitive information of the project from being leaked, achieving efficient audit supervision and enhancing the transparency and credibility of the management of the bank system research and development project.

[0033] 5. Task coordination and resource allocation phase The task coordination module deeply processes the requirement document. The requirement analysis unit uses an improved BERT model to train on the business logic and professional terminology in the bank requirement document, thereby accurately identifying the functional points, non-functional constraints, and associated business rules in the requirements, providing precise semantic understanding for subsequent task decomposition. The task decomposition unit uses the hierarchical task network (HTN) planning algorithm based on the hierarchical structure of the bank business process to gradually decompose the parsed requirements into executable task trees. Each task node defines the input and output parameters, preconditions and postconditions, and estimated workload information, and uses a dependency graph to clearly show the complex dependencies between tasks, ensuring the rationality of the task execution order and avoiding project delays or resource waste due to incorrect task order. The path recommendation unit combines a knowledge base containing various bank business architecture templates, and according to the current project requirement characteristics and the bank's technology stack status, uses a reinforcement learning algorithm to recommend the optimal development path for the task tree. This algorithm will continuously optimize the recommendation strategy based on feedback data from historical project path execution effects to improve development efficiency and project quality, ensuring that the project can be completed successfully within the specified time and meet the bank's business requirements and technical standards. The resource analysis unit uses graph neural network technology to construct a resource occupation graph, representing human and equipment resources as graph nodes and task resource demand relationships as graph edges. By analyzing the propagation path of resource conflicts using graph convolutional neural networks (GCN), it can predict potential resource conflict ranges in advance, thereby intelligently generating resource allocation schemes. During scheme development, various factors such as personnel skill matching, equipment availability time windows are considered to achieve efficient resource utilization and reasonable allocation, improve overall project execution efficiency, and reduce project risks caused by resource shortages or conflicts. In the task coordination and resource allocation stage, the system also has a management function for outsourcing tasks. The resource analysis unit can evaluate the allocation of internal resources and also assess the resource matching degree of outsourcing teams. By analyzing factors such as skill level, available time window, past project experience, and cost-effectiveness, it ensures that the resource requirements of outsourcing tasks are accurately matched.

[0034] 6. Security control phase The security control module runs throughout the entire bank system R&D project management system, and guarantees the security of the system and the confidentiality of the data. When the personnel log in the system, the authentication unit first starts the multi-modal biometric recognition technology, fuses multiple biometric features such as fingerprints, faces, and voiceprints for identity verification. The deep learning feature extraction model is used to improve the accuracy and anti-forgery ability of biometric recognition, effectively preventing illegal personnel from bypassing system authentication through forged biometric features. After the biometric recognition passes, the system will also generate a dynamic verification code with high security by combining the time-synchronized one-time password (TOTP) dynamic token technology, and require the user to input it to complete the double verification process, further enhancing the security of sensitive operations and ensuring that only authorized legal users can access and operate the sensitive resources of the system. The access control unit uses software-defined security (SDS) technology to build a flexible virtualized security boundary. According to the permission management and authentication results, the network access strategy is dynamically adjusted, and the sensitive data transmission channel is protected by an encrypted tunnel to prevent data from being stolen or tampered with during transmission. At the same time, the intrusion detection system (IDS) is used to monitor the access behavior of the user in real time, and the machine learning anomaly detection algorithm is used to analyze the access request in depth, and timely discover and block abnormal access requests, such as frequent access from unknown IP addresses, batch data download that does not conform to the user's normal operation habits, and comprehensively guard the security of the system, prevent data leakage and malicious attack events, and ensure the stable operation of the bank system R&D project management system and the security and reliability of the business data.

[0035] 7. Visualization and decision support stage The visualization display unit obtains data from each business module in real time, renders and displays according to the preset visualization rules, and provides a visual and comprehensive project management view for managers. For project progress, the system uses a Gantt chart to visually present the comparison between task planning and actual execution, clearly displays the progress deviation of tasks through different colored bar charts, and highlights tasks with progress lag through a traffic light color warning mechanism, making it easy for managers to quickly locate problem areas and take timely measures to adjust and optimize. In terms of risk display, a risk heat map is constructed, and risks are dynamically colored based on key dimensions such as risk level and impact range. The color from light to dark intuitively reflects the severity of the risk, helping managers quickly focus on key risk points, develop risk response strategies in advance, and reduce the impact of risks on the project. For resource occupation, a Sankey diagram is used for display, with flowing lines intuitively presenting the allocation and flow of resources between different tasks and departments, helping managers clearly understand resource utilization efficiency and bottlenecks, and thus making reasonable resource allocation and optimization. For outsourcing risk control, the progress, resource occupation, and risk situation of outsourcing tasks are also displayed, providing managers with an intuitive view of outsourcing management and helping them quickly grasp the execution status of outsourcing tasks and adjust strategies in a timely manner. Through WebSocket real-time push technology, the display data can be updated in a timely manner, reflecting the latest status of project management, making it easy for managers to keep abreast of project dynamics and make scientific and reasonable decisions, ensuring the smooth progress of the bank system development project and timely delivery of high-quality system results.

[0036] 8. Continuous optimization phase According to the feedback information and data analysis results of the system, the project team continuously optimizes the execution process of the project. For example, based on the compliance check list and difference report generated by the intelligent management platform, the project documents are updated and improved in a timely manner to ensure that the project always meets regulatory requirements; referring to the risk level prediction results and allocation suggestion scheme provided by the risk prediction module, the project plan is dynamically adjusted, resources are reasonably arranged, and project risks are reduced; combined with the task execution and resource occupation analysis data in the task collaboration module, the task priority and resource allocation scheme are optimized to improve project execution efficiency and quality; by monitoring the execution of outsourcing tasks and changes in risks in real time, the project team can adjust outsourcing strategies in a timely manner, optimize the allocation and management of outsourcing tasks, and ensure the smooth execution of outsourcing tasks, reducing the impact of outsourcing risks on the overall project. Throughout the project cycle, monitoring, evaluation, feedback, and optimization are continuously repeated to form a continuous improvement closed-loop management mechanism, ensuring that the bank system development project can be efficiently, stably, and compliantly advanced, and ultimately successfully delivered, providing strong support for the digital transformation and business development of the bank.

[0037] Through the above implementation process, the bank system research and development project management system can give full play to the functional advantages of each module, realize all-around and fine management of the bank system research and development project, effectively improve the management efficiency and quality of the project, reduce the project risk, and guarantee the stable development and reliable delivery of the bank system.

[0038] In addition, it should be noted that the present application can be provided as a method, apparatus or computer program product. Therefore, the embodiments of the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including a computer-usable program of instructions) containing computer-usable program code.

[0039] The embodiments of the present application are described with reference to flowcharts and / or block diagrams of the method, terminal device (system) and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, embedded processor or other programmable data processing terminal device to produce a machine, so that the instructions executed by the computer or other programmable data processing terminal device produce a device implemented in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow(s) or block(s).

[0040] These computer program instructions can also be stored in a computer-readable memory capable of guiding a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including instruction apparatus, which implements the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow(s) or block(s). These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operation steps are performed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide a process for implementing the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that performs the functions specified in the flow(s) or block(s). These computer program instructions can also be loaded into a computer or other programmable data processing terminal device, so that a series of operation steps are performed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide a process for implementing the flowcharts and / or block diagrams.

[0041] It is also noted that, as used herein, the terms "first", "second", and the like, merely designate a relationship or order of importance and do not necessarily indicate any such actual relationship or order of importance. The terms "comprises", "comprising", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0042] Finally, it is noted that the foregoing description is of a preferred embodiment of the application, and that numerous changes and modifications can be made thereto without departing from the spirit and scope of the application as set forth in the appended claims.

Claims

1. A banking system R&D project management system, characterized by: include: The intelligent management platform connects to the regulatory database in real time and uses NLP technology to scan sensitive words in project documents, generate compliance checklists and discrepancy reports, mark high-risk tasks, and recommend corrective measures. The risk prediction module uses machine learning to analyze historical project data, predict the risk level of R&D cycle delays and resource overspending, trigger early warnings, and recommend deployment plans; The blockchain evidence storage and traceability module stores the hash values ​​and operation records of key node documents on the chain, solidifies the timestamp and responsible person, and supports regulators to quickly access the full process audit log through private keys; The task collaboration module uses NLP to parse requirement documents, automatically breaking them down into executable task trees, analyzing conflicts between manpower and equipment through resource occupancy maps, and intelligently generating deployment plans. The security control module implements hierarchical permissions based on roles and data sensitivity. Sensitive operations require dual verification using biometric authentication and dynamic tokens to prevent unauthorized access.

2. The banking system R&D project management system according to claim 1, characterized in that: The intelligent management platform includes a data collection unit, a policy analysis unit, a sensitive word scanning unit and a report generation unit; The data collection unit ensures that incremental data is captured within the first time window of regulatory data updates and caches data through the message queue Kafk to prevent data loss due to sudden increases in volume and ensure data collection continuity. The policy parsing unit is pre-trained on the complex structure and professional terminology of financial policy texts, and then fine-tuned based on the bank's own business characteristics. The model can accurately identify key elements in the policy and convert policy clauses into structured data, providing an accurate basis for subsequent compliance checks; The sensitive word scanning unit builds a sensitive word knowledge graph based on banking business scenarios, and dynamically evaluates the risk level of sensitive words to avoid misjudgments caused by simple keyword matching; the report generation unit compares compliance requirements, sensitive word risk assessment results and project documents, and uses preset rule templates to automatically generate a compliance checklist containing detailed differences and risk distribution heat maps, and customizes rectification plans for each high-risk task to assist project teams in responding to policy changes efficiently.

3. The banking system R&D project management system according to claim 1, characterized in that: The risk prediction module includes a data storage unit, a machine learning analysis unit, a risk assessment unit and an early warning unit; The data storage unit categorizes and stores the structured and unstructured parts of historical project data, converts unstructured data into analyzable feature vectors, and stores them together with structured data to facilitate rapid retrieval and batch reading during subsequent model training. The machine learning analysis unit captures the fluctuation patterns of the R&D cycle, fits the multi-dimensional characteristics of resource consumption, builds a hybrid model, and continuously optimizes the model hyperparameters through cross-validation to improve prediction accuracy; The risk assessment unit builds a risk assessment model based on a Bayesian network, integrating the risk factors of R&D cycle delays and resource overruns and their correlations into a probabilistic graphical model. It dynamically updates the prior probability based on the current project status and calculates the posterior probability of the risk level in real time, providing a quantitative basis for early warning. The early warning unit is responsible for real-time monitoring of project resource usage indicators and setting multi-level early warning thresholds. When the degree of resource conflict reaches the threshold, it immediately sends early warning information to the relevant responsible persons through the message push service, and calls the resource allocation algorithm. Based on the current resource usage map, it provides the project team with suggestions including manpower allocation and equipment sharing solutions to ensure the smooth progress of the project.

4. The banking system R&D project management system according to claim 1, characterized in that: The blockchain evidence storage and tracing module includes a hash calculation unit, a blockchain storage unit, and an audit log management unit; The hash calculation unit is responsible for performing multi-layer hash operations on key node documents to generate highly secure document fingerprints, ensuring that any minor tampering can be detected promptly. At the same time, it uses GPU acceleration technology to improve hash calculation efficiency to meet the performance requirements of high-frequency updates of bank project documents. The blockchain storage unit builds a proof-of-stake network involving regulators, bank management, and project teams. Each block records the hash value of key node documents, operation records, timestamps, and the digital signature of the responsible person. By automatically verifying the legitimacy of transactions, it ensures that the data on the chain is authentic, trustworthy, and cannot be tampered with, providing a solid foundation for subsequent audits. The audit log management unit is designed with an encrypted indexing mechanism to ensure that only regulatory agencies can quickly locate the required log information within the authorized scope. At the same time, it uses zero-knowledge proof technology to enable regulatory agencies to verify the integrity and authenticity of the logs while protecting sensitive project information, thereby achieving efficient audit supervision.

5. The banking system R&D project management system according to claim 1, characterized in that: The task collaboration module includes a demand parsing unit, a task disassembly unit, a path recommendation unit and a resource analysis unit; The requirements parsing unit conducts specialized training on common business logic and terminology found in bank requirements documents, accurately identifying functional points, non-functional constraints, and associated business rules within the requirements, providing precise semantic understanding for subsequent task decomposition. The task decomposition unit is responsible for gradually decomposing the parsed requirements into an executable task tree according to the bank's business process hierarchy. Each task node includes clear input and output parameters, pre- and post-conditions, and estimated workload. Dependency graphs are also used to clearly display complex dependencies between tasks, ensuring a reasonable order for task execution. The path recommendation unit builds a knowledge base containing various banking business architecture templates. Based on the current project requirements and the current status of the bank's technology stack, it uses reinforcement learning algorithms to recommend the optimal development path for the task tree. The resource analysis unit constructs a resource occupancy graph, treating human and equipment resources as graph nodes and the task demand relationship for resources as graph edges. It uses the graph convolutional neural network (GCN) to analyze the resource conflict propagation path, predict the potential conflict scope, intelligently generate deployment plans, and comprehensively consider factors such as personnel skill matching and equipment available time window to achieve efficient resource utilization.

6. The banking system R&D project management system according to claim 1, characterized in that: The security control module includes a rights management unit, an authentication unit and an access control unit; The permission management unit builds a multi-level permission system, sets access levels according to sensitivity, and combines dynamic permission adjustment strategies to update permission configurations in real time based on project stages and personnel position changes, ensuring accurate and adaptive permission allocation. The authentication unit uses multimodal biometric technology and combines it with dynamic token technology to generate a highly secure dynamic verification code. The dual verification mechanism jointly ensures the security of sensitive operations and prevents unauthorized operations by unauthorized personnel. The access control unit is responsible for building a virtualized security boundary, dynamically adjusting network access policies based on permission management and authentication results, encrypting tunnels to protect sensitive data transmission channels, using intrusion detection systems to monitor access behavior in real time, and applying machine learning anomaly detection algorithms to promptly detect and block abnormal access requests, thus safeguarding system security in all aspects.

7. The banking system R&D project management system according to claim 1, characterized in that: The modules exchange information via a data transmission interface. The data transmission interface adopts the RESTful API specification to ensure the simplicity and universality of inter-module calls. The API gateway is used for request routing, load balancing, and authentication management to ensure secure and reliable data interaction. In order to achieve data sharing and collaborative work, each module follows a unified data exchange format during data transmission, standardizes the definition of data fields in combination with the data dictionary, and implements asynchronous communication through message middleware to ensure the timeliness and integrity of data transmission in high-concurrency scenarios, and avoid blocking caused by dependencies between modules.

8. The banking system R&D project management system according to claim 1, characterized in that: The intelligent management platform is also integrated with the bank's internal project management system and human resources management system; By building a unified data integration layer, data extraction, conversion and loading can be realized to ensure data interconnection and interoperability.

9. The banking system R&D project management system according to claim 1, characterized in that: The system also includes a visualization display unit for integrating visualization components of different business dimensions in the form of plug-ins; For project progress display, Gantt chart is used to visually present the comparison between task plan and actual execution status, combined with traffic light color warning mechanism to highlight progress deviation; In terms of risk display, a risk heat map is constructed, which is dynamically colored according to the risk level and impact range, helping managers quickly locate key risk points; Resource usage is displayed using a Sankey diagram, which clearly presents the resource flow and allocation status. WebSocket real-time push technology ensures timely updating of displayed data, making it easier for managers to make decisions and monitor, identify problems in a timely manner, and take appropriate measures.

10. The implementation process of a banking system R&D project management system according to any one of claims 1 to 9, characterized in that: The following steps are involved: Initialize system parameters and configure the functions and permissions of each module, including connecting to the regulatory database account information, setting basic permissions for each role, and initializing blockchain nodes; Through real-time integration with the regulatory database, the intelligent management platform uses pre-set data collection rules and NLP algorithms to parse and scan for sensitive terms in policy updates. Based on the rule engine, compliance checklists and discrepancy reports are generated, and to-do notifications are sent to relevant project leaders via the system's built-in push notification service. The risk prediction module is used to assess project risks. The model loads historical project data for self-training and optimization, and outputs risk level prediction results in real time based on the current project progress. When the risk reaches the warning threshold, the warning process is triggered and a deployment suggestion plan based on the current resource status is provided; The blockchain evidence storage and traceability module stores key node documents, automatically calculates document hash values, and stores relevant information on the blockchain. Operation logs are also recorded to ensure that every key operation can be traced. The task collaboration module parses, decomposes, and recommends paths for requirement documents. Project managers confirm the recommended task tree and development path in the system, and assign tasks to specific executors. The resource analysis unit monitors resource usage in real time and automatically adjusts or prompts manual intervention if conflicts occur. The security control module implements security controls throughout the entire process. Biometric identification and dynamic token verification are performed when personnel log into the system, and authorization is re-authenticated when accessing sensitive resources. All operation requests are released after security policy checks to ensure system security. The visualization display unit obtains data from each business module in real time, renders and displays it according to preset visualization rules, and assists managers in making decisions; Based on the system's feedback and data analysis results, the project team continuously optimized the project execution process until the project was completed, achieving efficient management and smooth delivery of the banking system R&D project.

Images