Financial intelligent analysis management method and system based on big data

Abstract

The invention relates to the technical field of data processing, in particular to a financial intelligent analysis management method and system based on big data. A financial intelligent analysis and management method based on big data comprises the following steps: S1, acquiring the permission of a target user, acquiring available resources of the target user according to the permission of the target user, acquiring latest policy and rule information from a data source, and acquiring key policy and rule information by using an NLP; and S2, inputting the key policy and rule information into an updating rule model to obtain an updated RPA operation rule, and generating a financial statement by using RPA software according to the available resources of the target user and the updated RPA operation rule. According to the method, by introducing a natural language processing technology and updating a rule model, the defect that dynamic adjustment cannot be achieved when RPA software is used is overcome, the sequence is optimized through a dynamic priority calculation formula and a genetic algorithm, and the flexibility and timeliness of the system are improved.

Description

Technical Field

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

[0002] With the deepening of digitalization of financial management, the policies and regulations faced by enterprises are becoming increasingly complex, and traditional financial processing methods are difficult to meet the requirements of efficiency, accuracy and compliance. In recent years, the application of robotic process automation (RPA), big data analysis and artificial intelligence technologies in the financial field has gradually increased, and it has become a trend to realize the automation and intelligence of financial operations through these technologies. However, the existing financial analysis management system still has the following shortcomings: traditional financial systems require a lot of manual intervention when dealing with frequently changing policies and regulations, and it is difficult to quickly update rules, which affects the timeliness and accuracy of financial report generation; when processing interrupted financial tasks, the existing system cannot dynamically adjust the priority of tasks, resulting in unreasonable resource allocation, processing delays and other problems, affecting overall efficiency. Summary of the invention

[0003] In order to overcome the shortcomings of RPA, such as the inability to adjust dynamically and being blind when not running successfully, the present invention provides a financial intelligent analysis management method and system based on big data.

[0004] The technical solution is: a financial intelligent analysis management method based on big data, including the following steps: S1: Obtain the target user's permissions, and based on the target user's permissions, obtain the target user's available resources, and at the same time obtain the latest policy and rule information from the data source, and use NLP to obtain key policy and rule information; S2: Input the key policy and rule information into the updated rule model to obtain updated RPA operation rules, and use the RPA software to generate financial statements based on the target user's available resources and the updated RPA operation rules; S3: saving the financial statement data interrupted in the generation of the financial statement, adding it to the sequence to be regenerated, and using a dynamic priority calculation formula to perform scheduling optimization on the sequence to be regenerated to obtain a first scheduling sorting sequence; S4: Use a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence, and regenerate a financial statement based on the second scheduling sorting sequence and the target user authority.

[0005] Preferably, the obtaining of the target user's permissions, and based on the target user's permissions, obtaining the target user's available resources, while obtaining the latest policy and rule information from the data source, and using NLP to obtain key policy and rule information, includes: binding the target user's permissions to his or her position or role based on RBAC, using natural language processing technology to pre-process the captured text data to obtain key policy and rule information, using a pre-trained NLP model to parse the extracted policy and rule information, generating a structured rule representation, and converting it into a format understandable to the RPA system.

[0006] Preferably, the key policy and rule information is input into the update rule model to obtain updated RPA operation rules, and financial statements are generated using RPA software based on the target user's available resources and the updated RPA operation rules, including: constructing an update rule model based on the Transformer model, using historical policy and rule data and its corresponding RPA operation rules as training data, training the model to learn the mapping relationship between policy and rule changes and RPA operation rules, and when the new policy or rule information is parsed and converted, using the trained update rule model to generate updated RPA operation rules, wherein the cosine annealing learning rate scheduling strategy is used to dynamically adjust the learning rate, and the gradient is clipped with a fixed threshold.

[0007] Preferably, the method of generating financial statements using RPA software based on the available resources of the target user and the updated RPA operation rules includes: combining the amount of available resources of the target user with the updated RPA operation rules of the target user to obtain financial statements, and encrypting the system operation log through an asymmetric encryption algorithm to form an encrypted log; linking the hash value of each encrypted log with the hash value of the previous log to build a log chain; distributing the log chain in multiple blockchain nodes and using a consensus mechanism to verify the validity and consistency of the log; adding a timestamp to each log and providing an audit interface, wherein the log chain storage adopts a hash structure based on a Merkle tree; and the distributed storage adopts a PBFT consensus mechanism.

[0008] Preferably, the financial report data interrupted in the generation of the financial report is saved, added to the sequence to be regenerated, and the sequence to be regenerated is scheduled and optimized using a dynamic priority calculation formula to obtain a first scheduling sorting sequence, including: collecting dynamic attributes of financial tasks, including task initial importance, waiting time, dependency, recovery complexity and resource consumption, and inputting them into the dynamic priority calculation formula to obtain the first scheduling sorting sequence, wherein the dynamic priority calculation formula is: ; In the formula, is the priority of the task, is the initial importance of the task, is the waiting time of the task, is the dependency of the task, is the recovery complexity of the task, is the resource consumption required for the task, is the weight parameter.

[0009] Preferably, optimizing the first scheduling and sorting sequence using a genetic algorithm to obtain a second scheduling and sorting sequence, and regenerating the financial statement according to the second scheduling and sorting sequence and the target user permissions, includes: based on the first scheduling and sorting sequence, randomly generating a number of initial solutions as individuals in the population; evaluating using a fitness function, and according to the fitness value, using the roulette wheel selection method to select individuals as parents; performing a crossover operation on the parent individuals using the partially mapped crossover method to generate new offspring individuals, randomly selecting individuals from the offspring, and performing a mutation operation based on the task scheduling priority; using the elitist retention strategy to retain the individual with the highest fitness, and repeating the selection until a preset number of iteration times or a fitness threshold is reached, and selecting the individual with the highest fitness as the second scheduling and sorting sequence.

[0010] Preferably, the randomly generating a number of initial solutions as individuals in the population; evaluating using a fitness function, includes: where the fitness function formula is: ; In the formula, is the fitness value, is the average completion time of the task, is the resource utilization rate, is the priority default factor, is the weight parameter.

[0011] Preferably, the regenerating the financial statement according to the second scheduling and sorting sequence and the target user permissions, includes: re-obtaining the target user permissions to obtain the available resources of the target user, and regenerating the financial statement according to the order of the second scheduling and sorting sequence, adjusting the initial importance of the task with insufficient target user permissions in the second scheduling and sorting sequence using an adjustment formula, and recording the elimination times of the task, where the adjustment formula is: ; In the formula, is the initial importance of the task after adjustment, is the elimination times of the task, is the adjustment parameter.

[0012] Preferably, the use of a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence includes: using a genetic algorithm to optimize the first scheduling sorting sequence, before obtaining the second scheduling sorting sequence, generating the first scheduling sorting sequence based on the adjusted initial importance of the tasks, and deleting tasks with a number of eliminations greater than or equal to a preset threshold and notifying relevant personnel.

[0013] Preferably, a financial intelligent analysis and management system based on big data includes: User rights management module, which binds user rights and positions based on the RBAC model, parses user rights, and obtains user available resources; The NLP policy parsing module uses a pre-trained NLP model to parse the latest policy and rule information, extract key policy points and generate structured rules, and convert the parsed rules into a format that can be recognized by RPA; The RPA rule update module builds an update rule model based on the Transformer model, uses historical policy and rule data to train the model, and dynamically generates updated RPA operation rules; A financial report generation module, which generates financial reports using RPA software according to the target user's available resources and the updated RPA operation rules; A scheduling and sorting module is used to sort the tasks to be regenerated according to a dynamic priority calculation formula to obtain a first scheduling and sorting sequence, and optimize the first scheduling and sorting sequence according to a genetic algorithm to generate a final second scheduling and sorting sequence; The blockchain log management module uses an asymmetric encryption algorithm to encrypt operation logs; The system feedback and adjustment module adjusts the importance of tasks for which authority is insufficient, records the number of times tasks are eliminated, and optimizes the priority sorting of subsequent tasks.

[0014] Beneficial effects: 1. This invention introduces natural language processing technology to automatically parse the latest policies and rules and generate structured rule representations. Combined with the Transformer-based update rule model, it can achieve rapid adaptation of policies and regulations without a lot of manual intervention, thus improving the flexibility and timeliness of the system. 2. Use a dynamic priority calculation formula to reasonably calculate the task priority based on the dynamic attributes of the task, such as initial importance, waiting time, dependency, and recovery complexity, and generate an optimized scheduling sequence. When RPA fails to generate financial statements due to network problems or insufficient permissions, the task is restarted, significantly improving the system's resource utilization and task completion rate. 3. Use asymmetric encryption algorithms to encrypt operation logs, and build log chains in combination with blockchain technology to ensure the integrity and immutability of operation logs. Through timestamps and distributed storage, provide reliable audit interfaces to achieve efficient tracing and supervision of financial operations. 4. In case of insufficient authority or task interruption, use the adjustment formula to dynamically adjust the importance of the task and record the number of eliminations to provide reference data for subsequent task scheduling and improve the system's intelligent decision-making capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a financial intelligent analysis management method based on big data of the present invention; Figure 2 This is a flow chart of a financial intelligent analysis and management system based on big data of the present invention. DETAILED DESCRIPTION

[0016] 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.

[0017] Embodiment 1: A financial intelligence analysis management method based on big data, such as Figure 1 and Figure 2 As shown, the following steps are included: S1: Obtain the target user's permissions, and based on the target user's permissions, obtain the target user's available resources, and at the same time obtain the latest policy and rule information from the data source, and use NLP to obtain key policy and rule information; Based on RBAC, the target user's permissions are bound to their positions or roles. Natural language processing technology is used to pre-process the captured text data to obtain key policy and rule information. The pre-trained NLP model is used to parse the extracted policy and rule information, generate a structured rule representation, and convert it into a format that the RPA system can understand.

[0018] It should be noted that this method uses a role-based access control model to bind the target user's permissions to his position or role. For example, the financial manager of a company has audit permissions, while ordinary financial personnel only have report generation permissions. Through the RBAC model, the scope of user permissions can be clearly defined, and the system resources available to the target user, such as financial software modules and data source access permissions, can be dynamically screened according to the scope of permissions; data sources include but are not limited to government policy documents, industry specification documents, and tax law announcements, and these text data are captured using crawler technology or API interfaces; the captured text data is pre-processed by denoising, word segmentation, and entity recognition to filter out key content related to the target user's job responsibilities, and the policy and rule texts are parsed using a pre-trained model based on deep learning; for example: for the text "the value-added tax rate is adjusted from 13% to 12%", the pre-trained model can parse the core information: "value-added tax", "tax rate", "13%->12%", and express it as a structured rule.

[0019] S2: Input the key policy and rule information into the updated rule model to obtain updated RPA operation rules, and use the RPA software to generate financial statements based on the target user's available resources and the updated RPA operation rules; Construct an update rule model based on the Transformer model, use historical policy and rule data and their corresponding RPA operation rules as training data, train the model to learn the mapping relationship between policy and rule changes and RPA operation rules, and when the new policy or rule information is parsed and converted, use the trained update rule model to generate updated RPA operation rules, in which the cosine annealing learning rate scheduling strategy is used to dynamically adjust the learning rate, and the gradient is clipped with a fixed threshold.

[0020] It should be noted that the Transformer model is used as the basic framework to build an update rule model. Due to its multi-head self-attention mechanism, Transformer can capture the complex relationship between policies and rule texts, and is suitable for processing rule generation tasks; historical policy and rule data and their corresponding RPA operation rules are used as training data to train the model to learn the mapping relationship between policy changes and RPA operation rules; when new policy or rule information is parsed, it is input into the trained update rule model to generate updated RPA operation rules; the cosine annealing learning rate scheduling strategy is used to dynamically adjust the learning rate to improve the convergence speed and stability of model training; according to the updated RPA operation rules, combined with the user's available resources, financial statements are automatically generated, and the gradient is clipped with a fixed threshold to prevent gradient explosion and improve the robustness of the training process.

[0021] The target user's available resource quantity is combined with the target user's updated RPA operation rules to obtain a financial statement, and the system operation log is encrypted by an asymmetric encryption algorithm to form an encrypted log; the hash value of each encrypted log is linked to the hash value of the previous log to build a log chain; the log chain is distributed and stored in multiple blockchain nodes, and a consensus mechanism is used to verify the validity and consistency of the log; a timestamp is added to each log and an audit interface is provided, wherein the log chain storage adopts a hash structure based on a Merkle tree; the distributed storage adopts a PBFT consensus mechanism.

[0022] It should be noted that the available resource quantity of the target user (such as computing power and storage space) is combined with the updated RPA operation rules, and the corresponding financial statements are generated according to the rules. The system operation log records the key events in the RPA operation process and encrypts the log using an asymmetric encryption algorithm to ensure that sensitive information can only be decrypted and read by authorized parties. The hash value of each encrypted log is linked to the hash value of the previous log to form an unalterable log chain. The log chain is stored in multiple blockchain nodes using a hash structure based on the Merkle tree to ensure the integrity of the data structure and verification efficiency. The PBFT consensus mechanism is used to verify the validity and consistency of the log. Even if some nodes fail, the system can still reach a consensus. Each log is timestamped to record the time information of the operation, which is convenient for auditing and tracing. An audit interface is provided to allow authorized users to query the operation records in the log chain.

[0023] S3: saving the financial statement data interrupted in the generation of the financial statement, adding it to the sequence to be regenerated, and using a dynamic priority calculation formula to perform scheduling optimization on the sequence to be regenerated to obtain a first scheduling sorting sequence; The dynamic attributes of financial tasks, including initial importance, waiting time, dependency, recovery complexity, and resource consumption of tasks, are collected and input into the dynamic priority calculation formula to obtain the first scheduling sorting sequence, where the dynamic priority calculation formula is: ; In the formula, is the priority of the task, is the initial importance of the task, is the waiting time of the task, For task dependencies, is the recovery complexity of the task, The resource consumption required for the task, is the weight parameter.

[0024] It should be noted that is the basic weight of the task, The waiting time from task creation to the current time. Whether the task depends on other tasks to complete, Regenerate the required resources and computational complexity for the task, The amount of computing resources required for task execution is calculated, and the sequence is regenerated according to the calculated priority to generate the first scheduling sorting sequence. The introduction of the dynamic priority formula can adjust the priority according to the actual attributes of the task, avoid unreasonable resource allocation, and improve task scheduling efficiency and system response speed.

[0025] S4: Use a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence, and regenerate a financial statement based on the second scheduling sorting sequence and the target user authority.

[0026] Based on the first scheduling sorting sequence, several initial solutions are randomly generated as individuals in the population; the fitness function is used for evaluation, and the roulette wheel selection method is used according to the fitness value to select individuals as the parent generation; the partial mapping crossover method is used to perform a crossover operation on the parent individuals to generate new offspring individuals, and individuals in the offspring are randomly selected to perform a mutation operation based on the task scheduling priority; the elite retention strategy is used to retain the individuals with the highest fitness, and the selection is repeated until a preset number of iterations or a fitness threshold is reached, and the individual with the highest fitness is selected as the second scheduling sorting sequence.

[0027] It should be noted that, based on the first scheduling sorting sequence, several initial solutions are randomly generated, each solution represents a task scheduling order, the number of individuals in the population is usually set according to the scale of the problem, and the fitness value is calculated for each individual in the population. The higher the fitness value, the better the task scheduling order of the individual. The roulette wheel selection method is used according to the fitness value, and individuals with high fitness values ​​are randomly selected from the population as parents. Partial mapping crossover operations are performed on the parent individuals to generate new child individuals. Individuals in the child generations are randomly selected, and their task scheduling order is mutated, and the order of certain tasks is adjusted, thereby increasing the diversity of the population and preventing the algorithm from falling into a local optimal solution. In each generation of iteration, the individual with the highest fitness value in the current population is retained to avoid the optimal solution being lost during the iteration process. The above operation is repeated until the preset number of iterations or fitness threshold is reached. Finally, the individual with the highest fitness value is selected as the second scheduling sorting sequence, and the financial statements are regenerated according to the second scheduling sorting sequence and the target user's authority, and the tasks with insufficient authority are adjusted and the number of eliminations is recorded.

[0028] The fitness function formula is: ; In the formula, is the fitness value, is the average task completion time, is the resource utilization rate, is the priority default factor, is the weight parameter.

[0029] It should be noted that the average completion time is an important indicator for evaluating scheduling efficiency, reflecting the time cost of task execution. Resource utilization reflects the efficiency of system resource use. Resource waste will reduce individual fitness. The priority default factor measures the consistency between the scheduling order and the expected priority to ensure that important tasks are completed first.

[0030] Reacquire the target user's authority to obtain the target user's available resources, and regenerate the financial report according to the order of the second scheduling sorting sequence. For the tasks in the second scheduling sorting sequence for which the target user's authority is insufficient, use the adjustment formula to adjust the initial importance of the task, and record the number of eliminations of the task, where the adjustment formula is: ; In the formula, is the initial importance of the task after adjustment, is the number of eliminations of the task, To adjust the parameters.

[0031] It should be noted that the system dynamically checks the target user's permissions and obtains the user's available resource list. This process ensures that when generating financial reports, the execution of tasks is subject to user permissions; each task is checked and attempted to be executed in turn according to the task execution order of the second scheduling sorting sequence to generate financial reports; the system maintains a counter for the number of eliminations for each task, recording the number of non-executions due to insufficient permissions, providing a basis for subsequent scheduling optimization and task allocation.

[0032] The first scheduling sorting sequence is optimized using a genetic algorithm. Before obtaining the second scheduling sorting sequence, the first scheduling sorting sequence is generated according to the adjusted initial importance of the tasks, and the tasks whose elimination times are greater than or equal to a preset threshold are deleted and relevant personnel are notified.

[0033] It should be noted that all tasks are re-sorted according to the adjusted initial importance to generate a new first scheduling sort sequence. The higher the importance of the task, the higher the priority. Tasks with a number of eliminations greater than or equal to the preset threshold (such as 5 times) are directly deleted from the task list, and the deletion log is recorded. At the same time, relevant personnel are notified for approval or reallocation.

[0034] Embodiment 2: Based on Embodiment 1, a financial intelligent analysis and management system based on big data includes: User rights management module, which binds user rights and positions based on the RBAC model, parses user rights, and obtains user available resources; The NLP policy parsing module uses a pre-trained NLP model to parse the latest policy and rule information, extract key policy points and generate structured rules, and convert the parsed rules into a format that can be recognized by RPA; The RPA rule update module builds an update rule model based on the Transformer model, uses historical policy and rule data to train the model, and dynamically generates updated RPA operation rules; A financial report generation module, which generates financial reports using RPA software according to the target user's available resources and the updated RPA operation rules; A scheduling and sorting module is used to sort the tasks to be regenerated according to a dynamic priority calculation formula to obtain a first scheduling and sorting sequence, and optimize the first scheduling and sorting sequence according to a genetic algorithm to generate a final second scheduling and sorting sequence; The blockchain log management module uses an asymmetric encryption algorithm to encrypt operation logs; The system feedback and adjustment module adjusts the importance of tasks for which authority is insufficient, records the number of times tasks are eliminated, and optimizes the priority sorting of subsequent tasks.

[0035] While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments.The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

Claims

1. A financial intelligence analysis and management method based on big data, characterized in that: S1: Obtain the target user's permissions, and based on the target user's permissions, obtain the target user's available resources, and at the same time obtain the latest policy and rule information from the data source, and use NLP to obtain key policy and rule information; S2: Input the key policy and rule information into the updated rule model to obtain updated RPA operation rules, and use the RPA software to generate financial statements based on the target user's available resources and the updated RPA operation rules; S3: saving the financial statement data interrupted in the generation of the financial statement, adding it to the sequence to be regenerated, and using a dynamic priority calculation formula to perform scheduling optimization on the sequence to be regenerated to obtain a first scheduling sorting sequence; S4: Use a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence, and regenerate a financial statement based on the second scheduling sorting sequence and the target user authority.

2. According to the financial intelligence analysis and management method based on big data according to claim 1, it is characterized in that: The method of obtaining the target user's permissions and, based on the target user's permissions, obtaining the target user's available resources, and obtaining the latest policy and rule information from the data source, and using NLP to obtain key policy and rule information includes: binding the target user's permissions to his or her position or role based on RBAC, using natural language processing technology to pre-process the captured text data to obtain key policy and rule information, using a pre-trained NLP model to parse the extracted policy and rule information to generate a structured rule representation, and converting it into a format that the RPA system can understand.

3. The financial intelligence analysis and management method based on big data according to claim 1 is characterized in that: The key policy and rule information is input into the update rule model to obtain the updated RPA operation rules, and the RPA software is used to generate financial statements according to the target user's available resources and the updated RPA operation rules, including: building an update rule model based on the Transformer model, using historical policy and rule data and its corresponding RPA operation rules as training data, training the model to learn the mapping relationship between policy and rule changes and RPA operation rules, and when the new policy or rule information is parsed and converted, using the trained update rule model to generate updated RPA operation rules, wherein the cosine annealing learning rate scheduling strategy is used to dynamically adjust the learning rate, and the gradient is clipped with a fixed threshold.

4. The financial intelligence analysis and management method based on big data according to claim 3 is characterized in that: The method of generating financial statements using RPA software according to the available resources of the target user and the updated RPA operation rules includes: combining the amount of available resources of the target user with the updated RPA operation rules of the target user to obtain financial statements, and encrypting the system operation log through an asymmetric encryption algorithm to form an encrypted log; linking the hash value of each encrypted log with the hash value of the previous log to build a log chain; distributing and storing the log chain in multiple blockchain nodes, and using a consensus mechanism to verify the validity and consistency of the log; adding a timestamp to each log and providing an audit interface, wherein the log chain storage adopts a hash structure based on a Merkle tree; and the distributed storage adopts a PBFT consensus mechanism.

5. The financial intelligence analysis and management method based on big data according to claim 1 is characterized in that: The method of saving the interrupted financial report data in the generation of the financial report, adding the data to the sequence to be regenerated, and using the dynamic priority calculation formula to schedule and optimize the sequence to be regenerated to obtain the first scheduling sorting sequence includes: collecting dynamic attributes of the financial task, including the initial importance, waiting time, dependency, recovery complexity and resource consumption of the task, and inputting the attributes into the dynamic priority calculation formula to obtain the first scheduling sorting sequence, wherein the dynamic priority calculation formula is: ； In the formula, MP is the priority of the task, P is the initial importance of the task, t is the waiting time of the task, D is the dependency of the task, C is the recovery complexity of the task, and R is the resource consumption required by the task. is the weight parameter.

6. The financial intelligence analysis and management method based on big data according to claim 1 is characterized in that: The method uses a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence, and regenerates financial statements based on the second scheduling sorting sequence and the target user authority, including: based on the first scheduling sorting sequence, randomly generating a number of initial solutions as individuals in the population; using a fitness function to evaluate, using a roulette wheel selection method according to the fitness value, and selecting individuals as parents; using a partial mapping crossover method to perform a crossover operation on the parent individuals to generate new offspring individuals, randomly selecting individuals from the offspring, and performing a mutation operation based on the task scheduling priority; using an elite retention strategy to retain the individuals with the highest fitness, repeating the selection until a preset number of iterations or a fitness threshold is reached, and selecting the individual with the highest fitness as the second scheduling sorting sequence.

7. The financial intelligence analysis and management method based on big data according to claim 6 is characterized in that: The method randomly generates a number of initial solutions as individuals in the population; and uses a fitness function for evaluation, including: wherein the fitness function formula is: ； In the formula, F is the fitness value, is the average task completion time, is the resource utilization rate, is the priority default factor, is the weight parameter.

8. The financial intelligence analysis and management method based on big data according to claim 7 is characterized in that: The regenerating financial statements according to the second scheduling sorting sequence and the target user's authority includes: reacquiring the target user's authority to obtain the target user's available resources, and regenerating financial statements according to the order of the second scheduling sorting sequence, adjusting the initial importance of the task for which the target user's authority is insufficient in the second scheduling sorting sequence using an adjustment formula, and recording the number of eliminations of the task, wherein the adjustment formula is: ； In the formula, N is the initial importance of the task after adjustment, k is the number of eliminations of the task, To adjust the parameters.

9. The financial intelligence analysis and management method based on big data according to claim 8 is characterized in that: The method of using a genetic algorithm to optimize the first scheduling sorting sequence to obtain a second scheduling sorting sequence includes: using a genetic algorithm to optimize the first scheduling sorting sequence, before obtaining the second scheduling sorting sequence, generating the first scheduling sorting sequence according to the adjusted initial importance of the tasks, and deleting tasks with a number of eliminations greater than or equal to a preset threshold and notifying relevant personnel.

10. A financial intelligent analysis management method based on big data, according to a financial intelligent analysis management system based on big data according to any one of claims 1 to 9, characterized in that: include: User rights management module, which binds user rights and positions based on the RBAC model, parses user rights, and obtains user available resources; The NLP policy parsing module uses a pre-trained NLP model to parse the latest policy and rule information, extract key policy points and generate structured rules, and convert the parsed rules into a format that can be recognized by RPA; The RPA rule update module builds an update rule model based on the Transformer model, uses historical policy and rule data to train the model, and dynamically generates updated RPA operation rules; A financial report generation module, which generates financial reports using RPA software according to the target user's available resources and the updated RPA operation rules; A scheduling and sorting module is used to sort the tasks to be regenerated according to a dynamic priority calculation formula to obtain a first scheduling and sorting sequence, and optimize the first scheduling and sorting sequence according to a genetic algorithm to generate a final second scheduling and sorting sequence; The blockchain log management module uses an asymmetric encryption algorithm to encrypt operation logs; The system feedback and adjustment module adjusts the importance of tasks for which authority is insufficient, records the number of times tasks are eliminated, and optimizes the priority sorting of subsequent tasks.

Images