RPA and knowledge graph-based financial whole-process collaborative management and control system

Abstract

This invention relates to the field of financial digital management technology, specifically a collaborative control system for the entire financial process based on RPA and knowledge graphs. The system includes an RPA data acquisition module, a data transmission module, an extraction and correlation analysis module, and a risk warning module. The RPA data acquisition module automates the collection, integrated management, and key information processing of the entire financial process to obtain financial process data. The data transmission module intelligently adjusts the data transmission path and obtains the target transmission path, then intelligently transmits the financial process data according to the target path to obtain a financial information database. The extraction and correlation analysis module performs data detection and correlation analysis on the financial information database to obtain a financial data knowledge graph. The risk warning module performs verification analysis and anomaly location on the financial data based on the financial data knowledge graph. This invention improves data management efficiency and achieves collaborative control and early warning for the entire financial process through automated processes and technologies.

Description

Technical Field

[0001] This invention relates to the field of financial digital management technology, specifically to a collaborative management and control system for the entire financial process based on RPA and knowledge graphs. Background Technology

[0002] Under the traditional financial management model, a large amount of repetitive and routine work not only requires a lot of time and energy, but is also prone to errors due to human negligence, which in turn seriously affects the efficiency and data quality of financial information management.

[0003] Financial data is often scattered across different systems and departments within a company. However, the lack of an effective data integration and sharing mechanism between different systems makes it difficult to transmit and integrate financial information in a timely and accurate manner, and makes it difficult to obtain comprehensive and accurate data when making decisions.

[0004] Traditional financial management models lack effective data analysis and monitoring methods, making it difficult to detect potential financial risks in a timely manner, such as abnormal fund flows and tax declaration errors. Traditional models mainly rely on existing experience and subjective judgment, lacking scientific data analysis models, which makes the decision-making process susceptible to personal factors and subjective biases, ultimately leading to inaccurate and unreasonable decision results.

[0005] Existing financial management systems have relatively limited functionality and are unable to meet the needs of collaborative management across the entire financial process. Some systems focus on financial accounting or report generation, lacking financial data analysis and risk warning functions; while other systems have certain data analysis capabilities, their data integration and sharing capabilities are weak, making it impossible to achieve effective integration of finance and business.

[0006] Therefore, there is a need for a system that can improve the efficiency of financial management, integrate financial data, enhance risk control capabilities, provide data support for financial system management decisions, and promote the development of financial management systems towards digitalization and intelligence. Summary of the Invention

[0007] To address the shortcomings of existing methods and the needs of practical applications, and to effectively solve the problems existing in the traditional financial management model, this system aims to comprehensively improve the data management efficiency of the system. Through automated processes and intelligent technologies, it reduces manual intervention and achieves efficient operation of financial work; strengthens data integration capabilities to ensure the comprehensiveness, accuracy, and timeliness of financial data; enhances risk control capabilities, analyzes potential risks in real time and provides early warnings, safeguards corporate financial security, and provides technical support for corporate management. On one hand, this invention provides a collaborative management and control system for the entire financial process based on RPA and knowledge graphs. The system includes: an RPA data acquisition module, a data transmission module, an extraction and correlation analysis module, and a risk warning module. The RPA data acquisition module automatically collects, manages, and processes key information from the entire financial process to obtain financial process data. The data transmission module, connected to the RPA data acquisition module, receives the financial process data and obtains a target transmission path based on path latency, weighting factors, and path load rate. It then transmits the financial process data according to the target transmission path to output a financial information database. The extraction and correlation analysis module, connected to the data transmission module, receives the financial information database and processes the data to output a financial data knowledge graph. The risk warning module, connected to the extraction and correlation analysis module, receives the financial data knowledge graph and performs risk analysis based on it to output security warning information, thus achieving collaborative management and security warning for the entire financial process.

[0008] The four modules of this invention work together to form a complete financial data processing chain, which closely connects all aspects of the financial process, from data collection, transmission, analysis to risk warning, realizing the automation and intelligent management of the financial process and improving the synergy and efficiency of financial management.

[0009] Optionally, the step of automatically collecting, integrating, and processing key information of the entire financial process through the RPA data acquisition module to obtain financial process data information includes: setting up an automated data acquisition mechanism in the RPA data acquisition module; simulating repetitive operations through the automated data acquisition mechanism, including invoice entry, data capture, and report generation; and automatically collecting the entire financial process information based on the automated data acquisition mechanism to obtain initial financial process information.

[0010] The automated data collection mechanism of this invention ensures that all data is collected in accordance with unified standards and formats, thereby improving the quality and comparability of financial data.

[0011] Optionally, the step of automatically collecting, integrating, managing, and processing key information of the entire financial process through the RPA data acquisition module to obtain financial process data information includes: setting up an integrated data management scheme in the RPA data acquisition module; detecting and adjusting the coverage of the initial financial process information according to the integrated data management scheme to obtain adjusted financial process information; setting up a key link processing mechanism in the RPA data acquisition module based on historical data and incident information; and supplementing and marking the adjusted financial process information in conjunction with the key link processing mechanism to obtain financial process data information.

[0012] The RPA data acquisition module of this invention can ensure that the collected financial process data follows a unified standard and format, making data sharing between different departments and systems easier and more efficient, and promoting collaborative work between different departments.

[0013] Optionally, the data transmission module is connected to the RPA data acquisition module to receive the financial process data information and obtain the target transmission path based on path delay, weight factor, and path load rate, including: obtaining data transmission path parameters according to data transmission optimization technology, the data transmission path parameters including path delay, weight factor, and path load rate; obtaining the communication path between edge nodes and aggregation nodes based on the data transmission module; obtaining link delay, decision factor, and congestion coefficient based on the communication path; and obtaining path delay by combining the communication path, the link delay, the decision factor, and the congestion coefficient.

[0014] This invention combines communication path, link delay, decision factor and congestion coefficient to obtain path delay, and the data transmission module can reasonably allocate bandwidth resources according to path delay and other parameters to further improve data transmission speed.

[0015] Optionally, the data transmission module is connected to the RPA data acquisition module to receive the financial process data information and obtain the target transmission path based on path delay, weight factor and path load rate, including: setting weight factors based on the data transmission module, wherein the weight factors include a first weight factor for adjusting path delay, a second weight factor for adjusting the degree of influence of load rate and a third weight factor for controlling the weight of path hop count.

[0016] The data transmission module of this invention adjusts the path and selection strategy according to the network topology, which is beneficial for selecting the optimal path for subsequent financial data transmission.

[0017] Optionally, the data transmission module is connected to the RPA data acquisition module to receive the financial process data information and obtain the target transmission path based on path latency, weighting factor, and path load rate, including: obtaining cache queue information, CPU utilization data, memory utilization information, and bandwidth utilization data from the data transmission module; analyzing the cache queue information, CPU utilization data, memory utilization information, and bandwidth utilization data, and determining the path load rate.

[0018] This invention reflects the resource consumption of each node on the path from multiple key dimensions such as storage, computing, memory and network, which can more comprehensively and accurately assess the load of the path and avoid misjudgment of the path load due to the deviation of a single indicator.

[0019] Optionally, the data transmission module is connected to the RPA data acquisition module to receive the financial process data information and obtain the target transmission path based on path delay, weight factor and path load rate, including: constructing a data transmission path objective function based on data transmission optimization technology, the path delay, the weight factor and the path load rate; the data transmission module uses the data transmission path objective function to intelligently adjust the data transmission path to obtain the target transmission path.

[0020] The objective function of this invention comprehensively considers multiple key factors such as path delay, weighting factor and path load rate, which can more accurately select the path with the highest transmission efficiency and significantly improve the overall data transmission efficiency of the system.

[0021] Optionally, the data transmission module is connected to the RPA data acquisition module to receive the financial process data information and obtain the target transmission path based on path delay, weight factor, and path load rate, including: randomly obtaining an initial path scheme based on chaotic mapping; evaluating the initial path scheme through the data transmission path objective function to obtain an evaluation result; selecting a transmission path with higher fitness based on the evaluation result to generate a new path; and performing crossover and mutation operations on the new path until the termination condition is met or the maximum number of iterations is reached to obtain the target transmission path.

[0022] The data transmission path objective function of this invention comprehensively considers multiple key factors such as path delay, weighting factor, and path load rate, enabling a comprehensive and accurate evaluation of path performance and improving the system's global search capability.

[0023] Optionally, the extraction and correlation analysis module is connected to the data transmission module and is used to receive the financial information database and process the data to output a financial data knowledge graph, including: setting a financial data detection mechanism in the extraction and correlation analysis module; performing time synchronization analysis, data verification, and data repair on the financial information database through the financial data detection mechanism, and outputting a detected financial information database; establishing a correlation analysis mechanism in the extraction and correlation analysis module; and using the correlation analysis mechanism to analyze and evaluate the detected financial information database to obtain a financial data knowledge graph.

[0024] This invention uses a financial data knowledge graph to display financial data and relationships, which can provide a more intuitive understanding of a company's financial status and business operations.

[0025] Optionally, the risk warning module is connected to the extraction and correlation analysis module, and is used to receive the financial data knowledge graph, and perform risk analysis based on the financial data knowledge graph to output security warning information, thereby realizing collaborative control and security warning of the entire financial process. This includes: extracting financial information to be monitored from the financial data knowledge graph; the risk warning module performing verification analysis on the financial information to be monitored to obtain tax difference analysis results and consistency verification results; performing data anomaly location based on the financial data knowledge graph, the tax difference analysis results, and the consistency verification results to obtain anomaly location results; combining anomaly risk classification information, the tax difference analysis results, the consistency verification results, and the anomaly location results to determine the risk level of the financial information to be monitored, and obtaining risk level determination results; and triggering a security warning processing flow based on the risk level determination results to achieve collaborative control and security warning of the entire financial process.

[0026] This invention combines abnormal risk classification information, tax difference analysis results, consistency verification results, and anomaly location results to determine the risk level of the financial information to be monitored. It can comprehensively consider the impact of multiple factors on financial risk and more accurately assess and predict risks in financial processes. Attached Figure Description

[0027] Figure 1 This is a flowchart of the financial end-to-end collaborative management and control system based on RPA and knowledge graphs according to the present invention.

[0028] Figure 2 This is a structural diagram of the financial end-to-end collaborative management and control system based on RPA and knowledge graph of the present invention. Detailed Implementation

[0029] Specific embodiments of the present invention will now be described in detail. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other instances, well-known circuits, software, or methods have not been specifically described to avoid obscuring the invention.

[0030] Throughout this specification, references to "an embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale.

[0031] Please see Figure 1 To improve financial information management, this invention introduces automated processes and intelligent technologies to reduce manual intervention and achieve automated and intelligent operation of financial workflows. It also strengthens data integration capabilities by centrally integrating and analyzing scattered financial data to provide data support for corporate decision-making. Furthermore, it utilizes advanced algorithms and models to enhance risk control capabilities, enabling real-time monitoring and analysis of financial data to ensure the safe and stable operation of the enterprise's financial management system. This invention proposes a collaborative management system for the entire financial process based on RPA and knowledge graphs. The system includes the following steps:

[0032] The financial end-to-end collaborative management system based on RPA and knowledge graphs includes: an RPA data acquisition module, a data transmission module, an extraction and correlation analysis module, and a risk warning module.

[0033] S1. The RPA data acquisition module automates the collection, integrated management, and key information processing of the entire financial process to obtain financial process data. The specific implementation steps are as follows:

[0034] In order to obtain comprehensive and accurate financial process data, this embodiment uses an RPA data acquisition module to automatically collect, manage and control, and process key information of the entire financial process.

[0035] The RPA data acquisition module incorporates an automated data acquisition mechanism. This mechanism simulates human operation to automate repetitive tasks such as invoice entry, data capture, and report generation. Combined with existing technologies and equipment, the automated data acquisition mechanism can collect multi-source information from different steps and operational stages of the financial process, ultimately obtaining the initial information of the financial process. The coverage and completeness of the aforementioned multi-source data acquisition play an important role in the analysis and management of the entire financial process.

[0036] The embodiments compare the operational efficiency of traditional manual data collection methods with that of automated data collection mechanisms.

[0037] Pre-set traditional manual data collection method to complete The total manual time required for each repetitive task (including invoice entry, data retrieval, report generation, etc.) is The unit is hours; the same data collection mechanism in the RPA data acquisition module is used to complete the same task. The total time required for each task is The unit is hours.

[0038] Based on the above information, it can be seen that the data collection efficiency improvement ratio of the automated data collection mechanism is [missing information]. The following relationship must be satisfied:

[0039] ,

[0040] in, This indicates the percentage improvement in data collection efficiency. Indicates the total manual time. This represents the total RPA time. It also represents the percentage increase in efficiency achieved through data acquisition. It is evident that using an RPA data acquisition module can significantly reduce the time required to complete repetitive tasks, improve the system's data acquisition efficiency, and save on the system's manpower and time costs.

[0041] Furthermore, the completeness of traditional manual data collection methods is compared with that of automated data collection mechanisms.

[0042] The theoretical number of data points to be collected for a certain operational stage in the entire financial process is set as follows: The actual number of data points collected through the automated data acquisition mechanism of the RPA data acquisition module is: .

[0043] Therefore, the completeness of data collection in this stage The following relationship must be satisfied:

[0044] ,

[0045] in, Indicates the completeness of multi-source data collection. This indicates the number of data points actually collected by the automated data collection mechanism. This indicates the theoretical number of data points that should be collected. Based on the completeness comparison results, it can be seen that the automated data collection mechanism can collect data more comprehensively, reduce data gaps, and ensure data completeness, providing a more reliable data foundation for subsequent financial analysis and process management.

[0046] An integrated data management and control scheme is implemented in the RPA data acquisition module. Based on this scheme, the coverage of initial financial process information is checked and adjusted to obtain the adjusted financial process information. In this embodiment, the integrated data management and control scheme combines edge computing with data aggregation technology, which can effectively improve the coverage and completeness of multi-source data acquisition. It can collect data from more sources and of a wider range of types, reduce data blind spots, and enable the system to grasp more comprehensive financial information.

[0047] This embodiment compares the coverage of traditional manual data collection methods with that of an integrated data management and control solution.

[0048] In the entire financial data management process, the traditional manual data collection method can cover the proportion of data collected. The integrated data management and control solution can cover a proportion of the collected data range. .

[0049] Based on the above two data range ratios, further analysis was conducted on the data collection coverage improvement ratio. And satisfy the following relationship:

[0050] ,

[0051] in, This indicates the percentage increase in data collection coverage. This indicates the proportion of data collected using traditional manual methods. This indicates the percentage of data collected that is covered by the integrated management and control data collection solution.

[0052] The above quantitative evaluation methods can clearly demonstrate the advantages of the integrated data management and control solution over traditional manual data collection methods, providing a clear basis for system optimization and improvement. By measuring the percentage increase in data collection coverage, we can intuitively understand the improvement in data collection coverage and thus better evaluate the system's performance and effectiveness.

[0053] Based on historical data and accident information, a key process processing mechanism is set up in the RPA data acquisition module. The adjusted financial process information is supplemented and marked in conjunction with the key process processing mechanism to obtain financial process data information.

[0054] This embodiment primarily relies on historical financial process data and incident records to identify key steps and error-prone areas within the entire financial process. Subsequently, for these steps, additional information collection and computing resources are allocated to process and analyze the generated data in real time. Data aggregation technology is then used to collect and supplement the adjusted financial process information, while simultaneously marking the data's origin. This process effectively reduces data transmission latency, accelerates response times, alleviates load on the system server, and ensures the timeliness of data collection throughout the financial process. Furthermore, this mechanism conserves network bandwidth and enhances data processing privacy and security.

[0055] The embodiments compare the data transmission between traditional manual data collection methods and key information collection and processing mechanisms.

[0056] The time delay from data collection to transmission to a remote data center using traditional manual methods is: The unit is seconds; the time delay for adopting the key link information collection and processing mechanism is... The unit is seconds.

[0057] Consequently, the system's data transmission latency is reduced by a certain percentage. The following relationship must be satisfied:

[0058] ,

[0059] in, Indicates the percentage reduction in data transmission latency. This indicates the time delay of traditional manual data collection methods. This indicates a time delay in the information collection and processing mechanism for key processes.

[0060] By combining the above steps with various comparative evaluation formulas, we can more intuitively and quantitatively evaluate the implementation effect and resource allocation of the RPA data acquisition module in the collection of financial process data information, providing technical support for the efficient management of the entire financial process.

[0061] S2. The data transmission module is connected to the RPA data acquisition module to receive financial process data information, obtain the target transmission path based on path delay, weight factor, and path load rate, and transmit the financial process data information according to the target transmission path to output a financial information database. The specific implementation details are as follows:

[0062] The data transmission module is closely connected to the RPA data acquisition module. This connection method establishes a basic channel for the efficient transmission of financial process data. The RPA data acquisition module can quickly acquire financial process data, which is then transmitted to the data transmission module in real time through the connection channel, providing a rich data source for subsequent data processing and transmission.

[0063] First, the data transmission module intelligently adjusts the data transmission path and obtains the target transmission path.

[0064] To achieve load balancing of financial data during transmission, this embodiment employs load balancing data transmission optimization technology. This technology can reasonably distribute financial data traffic to multiple servers or network nodes, which helps to achieve efficient transmission of financial data and improve the system's information collection capabilities.

[0065] In the data collection and transmission scenario of the entire financial process, there are several key links, including multi-source data collection, real-time processing and analysis, and data transmission. At different stages of the financial process, load balancing can further ensure the stability and timeliness of data transmission and the efficient operation of the system.

[0066] Based on the aforementioned data transmission optimization techniques, relevant parameters of the data transmission path can be obtained. In this embodiment, the data transmission path includes, but is not limited to, parameters such as path delay, weighting factor, and path load rate.

[0067] I. Path Delay

[0068] The data transmission module can clearly define the communication path between edge nodes and aggregation nodes. In this embodiment, symbols are used to represent this path. Represents edge nodes and aggregation nodes The communication path between them.

[0069] Based on the established communication path, further information such as link delay, decision factor, and congestion coefficient can be obtained. In the example, [the following is used]... Indicates link The path delay reflects the time required for data transmission on that link and is the foundational data for subsequent path delay calculations; Indicates link Decision factors on the link can determine the link Does it belong to a path? The specific judgment rule satisfies the following conditions:

[0070] ,

[0071] in, Indicates decision factors, Indicates a link. Represents edge nodes and aggregation nodes The communication path between them.

[0072] use Indicates link The network congestion coefficient, the value range of the above coefficient is: When the network is not congested As congestion increases The value also increases accordingly.

[0073] The path delay can be obtained by combining the communication path, link delay, decision factor, and congestion coefficient.

[0074] Path latency reflects the data's journey from edge nodes. via path Transmitted to the aggregation node The time delay experienced can measure data transmission efficiency. The formula for calculating path delay is as follows:

[0075] ,

[0076] in, Representing a path The time delay, Indicates a link. Represents edge nodes and aggregation nodes Communication path between them Indicates link The time delay on, Indicates link Decision factors on Indicates link The network congestion coefficient.

[0077] The above path delay calculation formula comprehensively considers multiple factors such as link delay, decision factors, and network congestion coefficient to evaluate and compare different transmission paths. This allows the system to select the path with the shortest delay for data transmission based on the calculated path delay value, avoiding the selection of paths with excessively long transmission times due to network congestion or poor link performance. This, in turn, ensures that financial data can be transmitted in a timely and efficient manner.

[0078] In a collaborative financial management system, a large amount of financial data needs to flow rapidly between different nodes and systems, via pathways. The latency calculation results allow the system to clearly understand the time required for data transmission, providing a quantitative basis for optimizing the transmission process and thus effectively improving data transmission efficiency.

[0079] II. Weighting Factors

[0080] In this embodiment, different weighting factors are set based on the data transmission module, which mainly include a first weighting factor for adjusting path delay, a second weighting factor for adjusting the degree of influence of load rate, and a third weighting factor for controlling the weight of path hop count.

[0081] The first weighting factor is the path delay adjustment factor.

[0082] First weighting factor The importance of path latency in the objective function can be adjusted. Path latency reflects the time it takes for data to travel from the source node to the target node. In actual business scenarios, different businesses have different sensitivities to latency. For financial transaction data transmission with high real-time requirements, latency is extremely sensitive, and the transmission time needs to be shortened as much as possible; while for some non-real-time financial statement data transmission, the latency requirement is relatively low.

[0083] First weighting factor The range of values ​​can be set as follows: First weighting factor The value needs to be adjusted and determined based on actual business requirements and the degree of latency sensitivity. To reduce the impact of latency on data transmission, the first weighting factor can be adjusted. Set it to a larger value; conversely, the first weighting factor can be appropriately reduced. The value of .

[0084] The second weighting factor is the load rate adjustment factor.

[0085] Second weighting factor It is mainly used to adjust the impact of load rate on the objective function. Load rate can reflect the busyness of network nodes or links. In the financial full-process collaborative management system, reasonable load balancing plays an important role in the stable operation of the system. If the load of any node or link is too high, it will lead to problems such as data transmission delay and packet loss, which will further affect the accuracy and timeliness of financial data.

[0086] Second weighting factor The range of values ​​can be set as follows: The relevant values ​​reflect the system's focus on load balancing. When the system prioritizes load balancing across nodes and links to avoid localized overload, the second weighting factor can be adjusted. Set it to a larger value; if the system has relatively low requirements for load balancing, the second weighting factor can be appropriately reduced. The value of .

[0087] The third weighting factor refers to the path hop count control factor.

[0088] Third weighting factor The weight of the path hop count in the objective function can be controlled. The path hop count refers to the number of intermediate nodes that data passes through from the source node to the target node. Based on practical applications, it is known that the more path hops there are, the greater the probability of failure or delay during data transmission.

[0089] Third weighting factor The range of values ​​can be set to Similarly, if the system can reduce intermediate steps in the data transmission process, the third weighting factor can be... Set to a larger value.

[0090] By reasonably setting the first weight factor Second weighting factor and the third weighting factor The data transmission module can weigh multiple indicators such as path latency, load rate, and path hop count according to actual business needs, which helps to select the optimal data transmission path and improve the data transmission efficiency and stability of the financial full-process collaborative management system.

[0091] III. Load Rate

[0092] First, the data transmission module acquires key data related to node load, including cache queue information, CPU utilization data, memory utilization information, and bandwidth utilization data. This data reflects the node's operating status and load level from different dimensions. Among them, cache queue information reflects the backlog of data currently being processed by the node; CPU utilization reflects the computing power occupied by the node's processing tasks; memory utilization shows the utilization of the node's storage resources; and bandwidth utilization indicates the node's busyness in network transmission.

[0093] To balance the impact of maximum load rate and load rate variance in load balancing, this embodiment introduces corresponding weight allocation factors. In practical applications, different scenarios have varying degrees of focus on different aspects of load balancing. If the business needs to avoid overloading individual nodes, the weight allocation factors can be tilted towards the maximum load rate. Different weight allocation factors can be flexibly adjusted according to actual business needs.

[0094] Further analysis of the maximum load rate was conducted, and the maximum load rate among all nodes was represented by... This means that during the path selection process, the node with the highest load rate needs to be fully considered. Based on the maximum load rate, nodes with excessive load can be effectively avoided, thereby achieving load balancing at the system level.

[0095] Load rate is mainly used to measure the load of nodes The load situation is analyzed in this embodiment, which combines cache queue information, CPU utilization data, memory utilization information, and bandwidth utilization data to determine the path load rate, and the node The load rate satisfies the following relationship:

[0096] ,

[0097] in, Indicates the load rate of different nodes. This represents the weight factor of the cache queue. Indicates the maximum cache queue length. This indicates the occupied cache queue of node i. Weighting factors representing CPU utilization This represents the CPU utilization of node i. Weighting factors representing memory utilization This represents the memory usage of node i. Weighting factors representing bandwidth utilization This indicates bandwidth utilization.

[0098] The maximum cache queue length is really a node The maximum amount of data a node can hold reflects its upper limit of processing capacity; The occupied cache queue refers to the current node The load rate of node i can be obtained by calculating the ratio of the occupied cache queue to the maximum cache queue length, based on the amount of data already stored in the cache.

[0099] To assess the load balancing across all nodes, the variance of the load rate of all nodes is further calculated, satisfying the following relationship:

[0100] ,

[0101] in, This represents the variance of the load rate across all nodes. Indicates the total number of nodes. Indicates the load rate of different nodes. This represents the average load rate of all nodes.

[0102] Other parameters.

[0103] The latency threshold of a data transmission path refers to a pre-defined standard value, which serves as an important criterion for measuring whether the path latency is within an acceptable range. For any path... Its delay is denoted as When the path delay When the latency exceeds the preset threshold, it indicates that the latency of this path is relatively large, which will affect the timeliness of data transmission, causing the data to fail to reach its destination within the specified time, and thus affecting the operating efficiency and performance of the entire system.

[0104] In the path Top edge node and aggregation nodes The hop count refers to the number of hops from the edge node. and aggregation nodes The number of intermediate nodes traversed during the data transmission process directly affects data transmission efficiency. In practical applications, too many hops will cause data to pass through more intermediate nodes during transmission, thereby increasing data transmission latency and reducing real-time performance.

[0105] In this embodiment, to ensure the efficiency and timeliness of data transmission, a hop count threshold is introduced as a parameter. The hop count threshold can limit the number of hops in a path to a reasonable range, avoiding the selection of paths with too many hops. When selecting a path, paths with more than the threshold will be excluded, thereby selecting paths with relatively fewer hops and higher transmission efficiency to ensure that financial data can be transmitted quickly and stably.

[0106] By following the steps above, we can comprehensively and accurately calculate the node load rate and system load rate variance, providing parameter information and data reference for the data transmission module to select the optimal path. This helps to achieve load balancing of the financial full-process collaborative management system and improve the system's stability and practicality.

[0107] Then, a data transmission path objective function is constructed based on data transmission optimization technology, path delay, weighting factors, and path load rate; the data transmission module uses the data transmission path objective function to intelligently adjust the data transmission path to obtain the target transmission path.

[0108] By combining data transmission optimization techniques and comprehensively considering parameters such as path latency, weighting factor, path load rate, and hop count, a data transmission path objective function is constructed. This objective function is mainly used to measure the comprehensive optimization target value of the data transmission path. By minimizing the target value, the optimal data transmission path is found to achieve load balancing and efficient transmission.

[0109] The objective function for the above data transmission path satisfies the following relationship:

[0110] ,

[0111] in, This represents the overall optimization target value for the data transmission path. This indicates the operation of finding the minimum value. Indicates the first weighting factor. Representing a path The time delay, Indicates the latency threshold of the data transmission path. This represents the second weighting factor. The weighting factor representing the maximum load factor and the load factor variance. This represents the maximum load rate among all nodes. Indicates the load rate of different nodes. This represents the variance of the load rate across all nodes. Indicates the third weighting factor. Representing a path Top edge node and aggregation nodes The number of jumps, Indicates the hop count threshold. Indicates the number of redundant paths. Indicates the actual number of redundant paths. This represents the theoretically optimal number of redundant paths.

[0112] The overall optimization objective value of the data transmission path is the value that is ultimately minimized by the entire objective function. It comprehensively reflects the impact of multiple factors such as path latency, load balancing, hop count, and redundancy on the performance of the data transmission path. By minimizing the objective value of the function, the data transmission path with the best overall performance can be found, thereby achieving load balancing and efficient transmission in the system.

[0113] The minimum value operation refers to selecting the path that minimizes the overall optimization objective value from all possible path options, i.e., the optimal data transmission path.

[0114] The number of redundant paths reflects the number of backup paths available for data transmission in a network. More redundant paths provide higher reliability and fault tolerance; when the primary path fails, it can quickly switch to a backup path, ensuring continuous data transmission.

[0115] The actual number of redundant paths refers to the number of redundant paths that are actually available in the current path scheme. By comparing it with the theoretically optimal number of redundant paths, the redundancy performance of the current path scheme can be evaluated.

[0116] In this embodiment, the theoretically optimal number of redundant paths is set as the reference standard for the number of redundant paths, which is used to measure whether the actual number of redundant paths has reached the optimal level. Considering the redundant path situation in the objective function helps to select paths with better reliability and fault tolerance.

[0117] In complex network topologies, to find the optimal data transmission path, this embodiment combines a genetic algorithm and a data transmission path objective function for path analysis and solution. The genetic algorithm can simulate natural selection and genetic mechanisms, and through simulating selection, crossover, and mutation operations in biological evolution, it iteratively optimizes the objective function to find the optimal solution that meets the conditions. The specific process is as follows:

[0118] The first step is to randomly obtain an initial path scheme based on chaotic mapping.

[0119] To enhance the diversity of path information and avoid the algorithm from getting trapped in local optima, the embodiment uses a chaotic mapping method to obtain the initial population. The chaotic mapping has randomness and ergodicity, and can randomly generate a set of initial path schemes, providing rich starting points for the subsequent optimization process.

[0120] The second step is to evaluate the initial path scheme using the data transmission path objective function to obtain the evaluation result.

[0121] The initial path schemes are comprehensively evaluated using a data transmission path objective function. This objective function considers multiple key parameters such as path delay, weighting factor, path load rate, and hop count, accurately measuring the overall optimization target value of the data transmission path. The evaluation results for each initial path scheme can be quickly obtained through calculation. Based on the evaluation results, individuals with higher fitness are selected for breeding; higher fitness indicates a better path scheme for that individual and a greater likelihood of becoming a candidate for the optimal solution.

[0122] The third step involves selecting a highly adaptive transmission path based on the evaluation results to generate a new path. The highly adaptive transmission path selected from the previous step is used as the parent path. A new path is then generated using existing simulation algorithms. This new path incorporates the excellent characteristics of the parent path and also exhibits better performance.

[0123] The fourth step involves performing crossover and mutation operations on the new path until the termination condition is met or the maximum number of iterations is reached, in order to obtain the target transmission path.

[0124] The newly generated paths are subjected to crossover and mutation operations to further enrich the population diversity. An adaptive crossover probability is further introduced into the above process. and mutation probability When population diversity is below a threshold, the adaptive crossover probability can be increased. to This enhances the algorithm's exploration capabilities, enabling it to discover more potential optimal paths; and reduces the mutation probability when approaching the optimal solution. to Based on the above adjustments, we can ensure stable convergence when approaching the optimal solution, thereby increasing the probability of finding the global optimum.

[0125] In this embodiment, the above selection, crossover and mutation processes need to be iterated until the termination condition is met. At the same time, taking into account the convergence speed and the business needs of the enterprise, the following two termination conditions are set.

[0126] The maximum number of iterations is set to 500. When the algorithm reaches 500 iterations, it stops iterating regardless of whether the optimal solution has been found. This can avoid excessive computation time for the data transmission module.

[0127] When the optimal solution shows no improvement or significant change for 20 consecutive iterations, it indicates that it has converged to a local or global optimum. At this point, the algorithm can be terminated. This approach balances computational cost and optimization effectiveness, ensuring that the data delivery module obtains good results within a reasonable timeframe.

[0128] In this embodiment, by setting the aforementioned iterative optimization method and termination conditions, the optimal data transmission path can be obtained. The data transmission module continuously optimizes and iterates the data transmission path using the data transmission path objective function, combined with key factors such as path latency and load rate. During the optimization process, by comprehensively considering multiple factors such as path latency, load balancing, and path hop count, the optimal data transmission path can be more accurately evaluated and selected, further ensuring the effective implementation of load balancing and efficient transmission. In practical applications, the weighting factors and parameters can be further adjusted and optimized according to specific business needs and data characteristics to better adapt to different network environments and business scenarios.

[0129] Finally, financial process data is intelligently transmitted according to the target transmission path to obtain a financial information database. In this embodiment, the financial process data output by the RPA data acquisition module is intelligently transmitted according to the aforementioned target transmission path, accurately and timely transmitting the relevant data to the data transmission module, thereby completing the entire financial process data acquisition and transmission work, and ultimately forming a financial information database. The above process ensures the efficient and secure transmission of financial data, while providing strong support for the company's financial management and decision-making.

[0130] S3, the extraction and correlation analysis module is connected to the data transmission module. It is used to receive the financial information database and process the data to output a financial data knowledge graph. The specific implementation details are as follows:

[0131] The connection between the extraction and correlation analysis module and the data transmission module further ensures the stable implementation of the financial information flow and processing. The data transmission module accurately and efficiently transmits the financial information database to the data transmission module, laying the foundation for subsequent data relationship analysis and detection. In this embodiment, establishing a transmission channel between the two modules ensures both the real-time nature of the data and meets the requirements for batch synchronous processing, thereby guaranteeing the consistency and accuracy of relevant information in the financial information database.

[0132] A financial data detection mechanism is set up in the extraction and correlation analysis module. This detection mechanism further ensures the data quality of the financial information database through time synchronization analysis, data verification and data repair, and outputs the detected financial information database.

[0133] Time synchronization analysis can achieve efficient data synchronization between ERP and financial systems and ensure data quality. This embodiment adopts a hybrid architecture of RESTful API and WebSocket, which can support real-time push when financial information is generated, and can also realize scheduled batch synchronization of inventory data in scenarios such as daily, weekly, monthly and quarterly.

[0134] A dynamically adjusted weighting factor is introduced based on matching business load and data collection time intervals. During peak business periods (such as month-end closing), it can be Weight reduced to To avoid false alarms; during off-peak hours The weight was adjusted to 1.0 to further ensure monitoring sensitivity.

[0135] In this embodiment, the interface response time monitoring formula satisfies the following relationship:

[0136] ,

[0137] in, Indicates the interface response time. This indicates that the weighting factor is dynamically adjusted. Indicates the total number of sampling requests. Indicates the request end timestamp. This indicates the start timestamp of the request.

[0138] Furthermore, a dynamic threshold for the interface response time was set. In this embodiment, predictions are primarily based on historical data, and dynamic settings are applied using the ARIMA model to balance system transmission efficiency and reliability. The system triggers performance alerts in real time, and through dynamic weight and threshold adjustments, it ensures the accuracy and timeliness of interface response time monitoring, avoiding false alarms or missed alarms caused by business fluctuations.

[0139] Data verification can ensure that financial data is formatted correctly and that its values ​​are reasonable.

[0140] The embodiment uses regular expressions and a business rule engine to validate the data format, including but not limited to checking the compliance of invoice numbers. The embodiment uses the Z-Score method to detect outliers in non-normally distributed data (invoice amounts), and the outlier detection function satisfies the following relationship:

[0141] ,

[0142] in, This indicates the outlier detection results. Represents the data point value. This represents the median. This indicates the absolute deviation from the median.

[0143] In an optional embodiment, for non-normally distributed data (invoice amounts), set By setting an anomaly threshold and marking relevant data as abnormal, a secondary review process is automatically triggered in conjunction with the business rules engine. Through the above format verification and anomaly detection, the accuracy and compliance of financial data can be ensured, providing a foundation for optional monitoring in the subsequent financial full-process collaborative management system.

[0144] Data repair: This mainly targets missing financial data. In this example, the moving average method is used to fill in the missing data, which can improve the integrity of the system data.

[0145] The above moving average formula satisfies the following relationship:

[0146] ,

[0147] in, Indicates the current time point Data prediction values, Indicates the weighting coefficient. Represents the actual values ​​for each historical period. Indicates the length of the window period. Indicates a time index.

[0148] The recommended window length can be dynamically adjusted according to the business cycle. "Time" refers to using data from the past three periods for predictive analysis.

[0149] The weighting coefficients reflect the degree of influence of historical financial data on the predicted data values, and can be either linearly decreasing weights or business rule weights.

[0150] In an optional embodiment, if the periodic purchase volume exceeds a threshold, the weighting coefficient can be increased to strengthen the influence of high-volume periods. Through weighted moving averages and exponential smoothing, missing purchase price data can be effectively filled in, improving data integrity and analytical reliability. Simultaneously, through a pre-developed API interface, real-time data synchronization between the ERP and financial systems is achieved, ensuring timely and accurate transmission of key information such as sales and purchase data to the financial system, providing a data foundation for tax filing.

[0151] An association analysis mechanism is established in the extraction and association analysis module. This mechanism is then used to analyze and evaluate the financial information database after testing, and a financial data knowledge graph is obtained.

[0152] An association analysis mechanism is established in the extraction and association analysis module. This mechanism is used to analyze and evaluate the detected financial information database, thereby obtaining a financial data knowledge graph.

[0153] In this embodiment, based on the data processed by the financial data detection mechanism, the correlation of data information is further analyzed in order to construct a high-quality financial data knowledge graph.

[0154] First, node feature analysis is performed. Entities such as invoices, contracts, and bank statements in the financial information database are coded with features. For example, invoice feature information includes amount, tax rate, invoice date (timestamp encoding), and supplier ID (embedded encoding); contract feature information includes contract amount, payment terms (30-day payment period mapped to values ​​0-1), and associated invoice ID.

[0155] In this embodiment, a graph attention network model is used to dynamically learn the weights between nodes using an attention mechanism, and to analyze related information such as payment terms, tax rates, and contract amounts in the financial information database.

[0156] The node embedding relationship can be represented as follows:

[0157] ,

[0158] in, Indicates node embedding. Representing a graph neural network, Represents the node feature matrix, This represents the adjacency matrix. Cross-validation is used to optimize model parameters, further ensuring its accuracy reaches [a certain level]. It can efficiently uncover potential relationships between entities, improving the accuracy and reliability of association rule mining.

[0159] The relevant content of quantitative indicators for association rule mining is as follows:

[0160] The accuracy formula satisfies the following relationship:

[0161] ,

[0162] in, Indicates the accuracy result. This represents the number of accurately mined association rules. This indicates the total number of association rules discovered.

[0163] This embodiment converts financial data into a transaction dataset, with each transaction representing a single transaction record. The dataset is scanned twice: first, frequent items are identified; second, a conditional pattern base is constructed, compressing storage space. Starting from the leaf nodes and tracing upwards, frequent itemsets are generated, and effective rules are selected based on support and confidence scores.

[0164] The support formula satisfies the following relationship:

[0165] ,

[0166] in, Indicates support level, This represents the number of transactions containing the itemset. This represents the total number of transactions.

[0167] The confidence level formula satisfies the following relationship:

[0168] ,

[0169] in, Indicates the confidence level. express Support express Support level. Filtering valid rules.

[0170] Next, a quality assessment of the knowledge graph is conducted. To ensure the quality of the financial data knowledge graph, this embodiment evaluates it from three aspects: structure, connectivity, and content.

[0171] Structural: Set thresholds based on business requirements, where the graph density formula satisfies the following: ,in The number of edges in the graph refers to the relationships or connections between nodes (business relationships such as contracts, associations, and invoices). The number of vertices in the graph refers to entities or objects (companies, departments, or individual business entities, when...). At that time, the graph density is allowed. This is to avoid over-connection. At the same time, higher weights are assigned to high-frequency relationships such as contracts, associations, and invoices, and resource allocation can be optimized by calculating node importance using the PageRank algorithm.

[0172] Connectivity: In this embodiment, a disjoint-set data structure algorithm is used for real-time calculation, when the proportion... The system triggers a graph fragmentation warning and automatically initiates the relationship completion process. At the same time, it uses the Dijkstra algorithm to calculate the shortest path and combines it with business rules to filter invalid paths, further improving query efficiency.

[0173] Content-based: The formula for calculating the entity attribute completeness rate is as follows: ,in Indicates the entity attribute completeness rate. This represents the number of entities with complete attributes. This represents the total number of entities, and missing attributes can be filled in using data repair logic.

[0174] In this embodiment, the extraction and association analysis module ensures data quality through time synchronization, data verification, and data repair via a financial data detection mechanism. At the same time, it combines graph neural networks and FP-Growth algorithms to analyze data associations, thereby obtaining a high-quality knowledge graph.

[0175] S4, the risk warning module is connected to the extraction and correlation analysis module. It is used to receive financial data knowledge graphs and perform risk analysis based on the financial data knowledge graphs to output security warning information, thereby realizing collaborative control and security warning of the entire financial process. The specific implementation content is as follows:

[0176] In this embodiment, the risk warning module and the extraction and correlation analysis module are connected. After the extraction and correlation analysis module transforms the financial information database into a financial data knowledge graph with a clear structure and close correlation, it transmits relevant information in a timely manner, so that the extraction and correlation analysis module can receive the above-mentioned financial data knowledge graph in real time and accurately, providing a comprehensive and reliable data foundation for subsequent risk analysis work.

[0177] The first step is to extract the financial information to be monitored from the financial data knowledge graph.

[0178] Financial information to be monitored is randomly extracted from the financial data knowledge graph. Through data extraction operations, relevant data such as contract amount and contract number of contract node, and invoice amount, invoice number, tax rate and tax difference calculation rules of invoice node are obtained.

[0179] In this embodiment, financial information is retrieved using the SPARQL query language. The specific query statement is as follows:

[0180] SELECT contract invoice contract_amount invoice_amount tax_rate

[0181] WHERE {

[0182] contract rdf:type :Contract;

[0183] :hasAmount contract_amount.

[0184] invoice rdf:type :Invoice;

[0185] :hasAmount invoice_amount;

[0186] :hasTaxRate tax_rate;

[0187] :isAssociatedWith contract.

[0188] }

[0189] The second step, the risk warning module, verifies and analyzes the financial information to be monitored, obtaining the tax difference analysis results and consistency verification results of the financial information to be monitored.

[0190] The data anomaly detection content is as follows:

[0191] Construct a tax difference calculation model that satisfies the following relationship: ,

[0192] in, Indicates the tax difference. Indicates the invoice amount. Indicates the discount factor. Indicates the tax rate. This represents the actual tax amount. The above calculation model introduces a discount factor to correct the tax difference calculation model. If the error is not found, it may be due to an incorrect invoice tax rate, improper application of discounts, or changes in tax policies. The actual tax amount can be obtained directly from the invoice node attributes or calculated and verified by multiplying the invoice amount by the tax rate.

[0193] Based on the above tax difference calculation results, a formula for verifying the consistency between the contract and the invoice is constructed, and the following relationship is satisfied:

[0194] ,

[0195] in, This represents the absolute value of the difference. Indicates the contract amount. Indicates the invoice amount. This indicates the tax difference. In this example, the contract amount refers to the total amount stipulated in the contract excluding tax; the invoice amount refers to the total amount stated on the invoice including tax; if... This indicates that the contract amount perfectly matches the sum of the invoice amount and the tax difference; if This indicates that there are problems such as incorrect invoice amount, incorrect tax difference calculation, or failure to synchronize contract changes.

[0196] Further set dynamic thresholds for the amount In the example, when the absolute value of the difference is greater than the dynamic threshold for the amount, i.e. An anomaly warning is triggered. In this example, the ARIMA model is used to predict dynamic thresholds based on historical data. The predictive analysis formula is as follows:

[0197] ,

[0198] in, Indicates dynamic threshold. This represents the mean of the absolute values ​​of the historical differences. Indicates the risk coefficient. It represents the standard deviation.

[0199] The above risk coefficient Primarily set based on the level of business risk. The larger, The higher the value, the more conservative the warning. In normal scenarios, it can be set to... At this point, a small deviation is allowed, suitable for high-frequency, low-amount transactions such as retail; in high-risk scenarios, it is set to... This application scenario requires strict control and is suitable for large contracts or highly regulated industries such as finance and pharmaceuticals.

[0200] The historical difference mean refers to all contract-invoice pairs over a past period (e.g., 12 months). The average value. In an optional embodiment, historical data is... ,at this time This mean reflects the average deviation level under normal business conditions and helps to distinguish between systematic errors and random anomalies.

[0201] Standard deviation is primarily used to measure the volatility of historical financial data, i.e., the dispersion of the deviation, and satisfies the following relationship: , A larger threshold indicates that the historical financial data has fluctuated significantly. In this case, the threshold needs to be increased to avoid false alarms.

[0202] The third step is to locate data anomalies based on the aforementioned financial data knowledge graph, tax difference analysis results, and consistency verification results. The anomaly location results can be obtained by querying the relationship paths of all contracts and invoices through the financial data knowledge graph, thereby realizing the location analysis of anomaly information.

[0203] The fourth step involves classifying the anomalies based on the aforementioned anomaly risk classification information, tax difference analysis results, consistency verification results, and anomaly location results. In this example, the classification is primarily based on the absolute value of the difference between the contract and invoice consistency verification results. and dynamic threshold The risk levels are classified, and the classification results are as follows;

[0204] Low risk: This indicates a low risk level, corresponding to a yellow alert.

[0205] Medium risk: This level of risk corresponds to an orange alert.

[0206] High risk: This is considered a high-risk situation and corresponds to a red alert.

[0207] Based on the above risk levels, the risk level of the financial information to be monitored can be determined, and the risk level judgment result can be obtained.

[0208] The fifth step involves triggering a security warning process based on the risk level assessment results, in order to achieve collaborative control and security warnings across the entire financial process.

[0209] When an anomaly is detected, the risk warning module automatically generates a review task through the rules engine and pushes it to the financial audit management system. Simultaneously, it sends structured warning information to smart communication devices such as WeChat and DingTalk, including but not limited to key information elements such as contract number, invoice number, absolute value of the difference, and risk level. Furthermore, it can also simultaneously send a detailed warning report to the overall financial management system, including anomaly data statistics, trend analysis charts, and handling suggestions.

[0210] You can also view the details of abnormal data through a visual interface, compare the scanned copies of contracts with electronic invoices, and finally feed the data verification results back to the system to update the status markers of the corresponding nodes in the knowledge graph as verified or pending processing.

[0211] The system risk warning module can automatically generate a visual report containing multi-dimensional information, including heatmaps of abnormal data distribution, risk trend prediction curves, contract management analysis charts, etc. At the same time, the warning results are synchronized to the system to trigger enhanced review rules in the contract approval process, serving as a reference for internal control optimization.

[0212] Through the above systematic and logically clear steps and formulas, the risk warning module can automate the entire process from data analysis, anomaly detection, warning triggering to processing and feedback of the financial information to be monitored, ensuring that financial data risks are detected in a timely manner and accurately controlled, providing a solid guarantee for the financial security of enterprises.

[0213] Please see Figure 2 In an optional embodiment, the present invention also provides a financial end-to-end collaborative management and control system based on RPA and knowledge graph. The system includes an RPA data acquisition module, a data transmission module, an extraction and correlation analysis module, and a risk warning module. The modules are interconnected and implement the specific steps of the relevant embodiments of the financial end-to-end collaborative management and control system based on RPA and knowledge graph provided by the present invention, thereby improving the overall applicability and practical application capability of the system of the present invention.

[0214] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A financial whole-process collaborative management and control system based on RPA and a knowledge graph, characterized in that, The system comprises an RPA data collection module, a data transmission module, an extraction and correlation analysis module, and a risk early warning module. The RPA data collection module is used to automatically collect, integrally control, and process key information of the whole financial process to obtain financial process data information. The data transmission module is connected with the RPA data collection module, used to receive the financial process data information, and obtain a target transmission path based on path time delay, weight factor, and path load rate, and transmit the financial process data information according to the target transmission path to output a financial information database. The extraction and correlation analysis module is connected with the data transmission module, used to receive the financial information database, and process the data to output a financial data knowledge graph. The risk early warning module is connected with the extraction and correlation analysis module, used to receive the financial data knowledge graph, and perform risk analysis based on the financial data knowledge graph to output safety early warning information, realizing collaborative control and safety early warning of the whole financial process. The data transmission module is connected with the RPA data collection module, used to receive the financial process data information, and obtain a target transmission path based on path time delay, weight factor, and path load rate, which comprises: obtaining data transmission path parameters according to data transmission optimization technology, the data transmission path parameters comprising path time delay, weight factor, and path load rate; obtaining a communication path between edge nodes and sink nodes based on the data transmission module; obtaining link time delay, decision factor, and congestion coefficient according to the communication path; obtaining path time delay in combination with the communication path, the link time delay, the decision factor, and the congestion coefficient; constructing a data transmission path target function in combination with the data transmission optimization technology, the path time delay, the weight factor, and the path load rate; the data transmission module intelligently adjusts the data transmission path using the data transmission path target function to obtain a target transmission path. The RPA data collection module is used to automatically collect, integrally control, and process key information of the whole financial process to obtain financial process data information, which comprises:

2. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, setting a data automatic collection mechanism in the RPA data collection module; simulating repetitive operations, including bill entry, data capture, and report generation, through the data automatic collection mechanism; automatically collecting financial process information based on the data automatic collection mechanism to obtain financial process initial information. The RPA data collection module is used to automatically collect, integrally control, and process key information of the whole financial process to obtain financial process data information, which comprises:

3. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 2, characterized in that, setting a data integration control scheme in the RPA data collection module; ​ According to the data integration control scheme, the coverage degree of the initial financial process information is detected and adjusted to obtain adjusted financial process information; Based on historical data and accident information, a key link processing mechanism is set in the RPA data collection module; The adjusted financial process information is supplemented and marked based on the key link processing mechanism to obtain financial process data information.

4. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, The data transmission module is connected with the RPA data collection module, used to receive the financial process data information, and based on path delay, weight factor and path load rate to obtain the target transmission path, including: Based on the data transmission module, a weight factor is set, including a first weight factor for adjusting path delay, a second weight factor for adjusting load rate influence degree, and a third weight factor for controlling path hop count weight.

5. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, The data transmission module is connected with the RPA data collection module, used to receive the financial process data information, and based on path delay, weight factor and path load rate to obtain the target transmission path, including: According to the data transmission module, cache queue information, CPU usage rate data, memory usage rate information and bandwidth usage rate data are obtained; The cache queue information, CPU usage rate data, memory usage rate information and bandwidth usage rate data are analyzed to determine the path load rate.

6. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, The data transmission module is connected with the RPA data collection module, used to receive the financial process data information, and based on path delay, weight factor and path load rate to obtain the target transmission path, including: Based on chaotic mapping, an initial path scheme is randomly obtained; The initial path scheme is evaluated by the data transmission path target function to obtain an evaluation result; According to the evaluation result, a transmission path with higher fitness is selected to generate a new path; The new path is subjected to cross and mutation operations until the termination condition is met or the maximum iteration number is reached to obtain the target transmission path.

7. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, The extraction and correlation analysis module is connected with the data transmission module, used to receive the financial information database, and the data is processed to output a financial data knowledge graph, including: A financial data detection mechanism is set in the extraction and correlation analysis module; The financial information database is subjected to time synchronization analysis, data verification and data repair by the financial data detection mechanism, and a detected financial information database is output; An association analysis mechanism is established in the extraction and correlation analysis module; The detected financial information database is analyzed and evaluated by the association analysis mechanism to obtain a financial data knowledge graph.

8. The RPA and knowledge graph-based financial whole-process collaborative management and control system according to claim 1, characterized in that, The risk early warning module is connected with the extraction and correlation analysis module, used to receive the financial data knowledge graph, and based on the financial data knowledge graph, risk analysis is performed to output security warning information, realizing collaborative control and security warning of the whole financial process, including: Extracting the financial information to be monitored from the financial data knowledge graph; The risk early warning module performs verification analysis on the financial information to be monitored to obtain tax difference analysis results and consistency verification results of the financial information to be monitored; According to the financial data knowledge graph, the tax difference analysis result and the consistency verification result, data anomaly positioning is performed to obtain an anomaly positioning result; In combination with the anomaly risk grading information, the tax difference analysis result, the consistency verification result and the anomaly positioning result, risk level judgment is performed on the to-be-monitored financial information, and a risk level judgment result is obtained; Based on the risk level judgment result, a security early warning processing flow is triggered to realize collaborative management and security early warning of the whole financial process.

Images