Enterprise financial data analysis method and system based on big data processing

Abstract

The application discloses an enterprise financial data analysis method and system based on big data processing, which comprises the following steps: collecting multi-source heterogeneous financial data of an enterprise, generating a global financial relationship structure and a partition structure set; performing hierarchical structure etching processing, generating an etching skeleton structure, and constraining the etching process; performing continuous structure etching evolution, generating an etching stable structure; injecting constraint driving and performing structure domestication processing, generating a domestication evolution structure; performing structure convergence behavior analysis, dividing structure domesticable states, and generating a structure domesticable state set; constructing a structure domestication adjustment mapping relationship, and generating a financial data analysis result. The application realizes financial data analysis through a structure etching and constraint propagation method, and has the advantages of clear structure expression and high analysis accuracy.

Description

Technical Field

[0001] This invention relates to the field of enterprise financial data analysis, and in particular to an enterprise financial data analysis method and system based on big data processing. Background Technology

[0002] As enterprises become more information-based, their financial data is characterized by being multi-sourced, heterogeneous, large-scale, and complex. Existing technologies typically process and analyze financial data through data integration, rule analysis, or statistical models. Common methods often rely on single modeling of business transaction data, cash flow data, or account relationship data, or on constructing simple relationship networks for correlation analysis, thereby enabling the monitoring and evaluation of the enterprise's financial situation. These methods can reflect the basic relationships between financial data to a certain extent, but overall they remain at the level of shallow structural analysis and static data processing.

[0003] When faced with complex financial relationship networks, existing technologies struggle to effectively identify redundant and deep critical connections. They lack dynamic analysis capabilities based on structural evolution, making analysis results susceptible to interference from noise. Furthermore, existing methods typically fail to constrain the structure by incorporating fund flow paths and transaction relationships, and lack mechanisms for state division and mapping adjustment based on structural convergence behavior. Consequently, they struggle to form a deep characterization and detailed analysis of the financial structure, thus limiting the accuracy and interpretability of financial data analysis results. Summary of the Invention

[0004] One objective of this invention is to propose a method and system for enterprise financial data analysis based on big data processing. This invention achieves financial data analysis through structural etch and constraint propagation methods, and has the advantages of clear structural expression and high analytical accuracy.

[0005] A method for analyzing enterprise financial data based on big data processing according to an embodiment of the present invention includes the following steps: Collect multi-source heterogeneous financial data from enterprises, including business transaction data, cash flow data, and account association data. Perform partitioning and parallel structured processing, perform relationship association and structure expression within each partition, and perform cross-partition fusion on the structure of each partition to generate a global financial relationship structure and a set of partition structures. A hierarchical structure etching process is performed on the global financial relationship structure. In each partition of the partition structure set, the connection relationship is weakened and redundant paths are stripped in parallel. The etching results of each partition are then aggregated globally to generate an etched skeleton structure. The etching process is constrained based on the capital flow path in the capital flow data and the transaction relationship in the business transaction data. Continuous structural etching evolution is performed on the etched framework structure. Deep connections are exposed and the structural core is stabilized through progressive etching operations, resulting in an etched stable structure. Constraint driving is injected into the etched stable structure and structural acclimatization processing is performed. The constraint driving comes from the account connection relationship in the account association data. The acclimatization evolution structure is generated through constraint propagation and structural convergence evolution. Perform structural convergence behavior analysis on the domestication and evolution structure, divide the structure into domesticable states based on the convergence behavior of the structure in the structural domestication process, and generate a set of domesticable states. Based on the set of structurally tamable states, a structural taming adjustment mapping relationship is constructed to generate financial data analysis results.

[0006] Optionally, the generation of the global financial relationship structure and partition structure set specifically includes: Collect multi-source heterogeneous financial data from enterprises, and perform unified identification and time alignment processing on the multi-source heterogeneous financial data to form relatable data under a unified time benchmark; Based on the data source attribute, business type attribute, and time interval attribute, the associative data is partitioned to generate a set of partition structures; Parallel structuring is performed within each partition of the partition structure set. Field parsing and data mapping are performed on the associative data within the partition, which are then converted into the corresponding structural expression form to generate the structure of each partition. Within the partition structure, relationship association processing is performed, transaction relationships between data are established based on business transaction data, and fund flow paths between data are established based on fund flow data, forming corresponding relationship connections within the partition structure; Introduce account-related data into each partition structure, perform connection relationship construction and expansion processing on the partition structure, embed account connection relationships into relationship connections and complete structural integration to form a partition structure with a consistent structure. Perform cross-partition fusion processing on the partition structure set, connect and integrate the partition structures based on the temporal continuity between partitions, generate a global financial relationship structure, and output the partition structure set synchronously.

[0007] Optionally, the generation of the etched skeleton structure specifically includes: Obtain the global financial relationship structure and partition structure set, perform hierarchical structure etching on the global financial relationship structure, expand the connection relationship according to the hierarchical relationship and perform layered action to generate a hierarchical connection relationship structure; Within each partition of the partition structure set, structural etching is carried out synchronously based on the hierarchical connection relationship structure. The connection relationships in each partition are filtered, and the connection relationships that appear repeatedly at the same level are regarded as redundant connection relationships. The redundant connection relationships are stripped to generate a weakened connection relationship structure. To weaken the connection structure, the redundant connection relationships are stripped along the connection path, and the connection relationships that maintain the connectivity of the partition structure are retained, thus generating the stripped partition structure. Based on the weakened connection structure and the partitioned structure after stripping, the fund flow path in the fund flow data is introduced to constrain the changes in the connection relationship. At the same time, the transaction relationship in the business transaction data is combined to limit the scope of the connection relationship to be retained, and a constrained partitioned structure is generated. The constrained partition structures within each partition are aggregated to form the corresponding partition etching results. The partitioned etching results are integrated as a whole, and the global aggregation is completed by restoring the connection relationship between the partitions to generate the etched skeleton structure.

[0008] Optionally, the generation of the etch-stabilized structure specifically includes: The connection relationships in the etched skeleton structure are etched layer by layer, and the connection relationships are screened out layer by layer along the connection path to generate the etched skeleton structure after layer-by-layer etching. In the etched skeleton structure after layered etching, the connection relationship is progressively etched, and the connection relationship is peeled off step by step according to the hierarchical order to generate the etched skeleton structure after progressive etching. In the etched skeleton structure after progressive etching, the connection relationships are screened, and the connection relationships retained after etching are extracted to generate deep connection relationships; The etching skeleton structure is connected and fixed based on deep connectivity relationships. The deep connectivity relationships are written into the etching skeleton structure to generate a stable connectivity structure. The stable connection structure is integrated as a whole to generate an etched stable structure.

[0009] Optionally, the generation of the domestication evolution structure specifically includes: Write the account connection relationships in the account association data into the connection relationships in the etched stable structure to form an etched stable structure containing the account connection relationships; Based on the etched stable structure containing account connection relationships, constraint propagation is performed on the connection relationships. The account connection relationships are passed level by level along the connection path to generate the etched stable structure after constraint propagation. In the etched stable structure after constraint propagation, the connection relationships are filtered, retaining those consistent with the account connection relationships and stripping those that do not conform to the account connection relationships, thus generating the etched stable structure after constraint filtering. Based on the etch-stable structure after constraint screening, the connection relationships are subjected to structural convergence evolution. The retained connection relationships are then aggregated along the connection path to generate the converged etch-stable structure. The converged etched stable structure is integrated as a whole to generate a domestication and evolution structure.

[0010] Optionally, the generation of the set of structurally tunable states specifically includes: The connection relationships in the domestication and evolution structure are expanded along the connection paths, and the connection paths and node positions of each connection relationship in the structure are recorded to generate the domestication and evolution structure after the connection paths are expanded. Based on the domestication evolution structure after the connection path is expanded, the structural convergence behavior of the connection relationship is analyzed. The retention and stripping of the connection relationship in the structural domestication process are identified along the connection path, and the domestication evolution structure corresponding to the structural convergence behavior is generated. In the domestication and evolution structure corresponding to the structural convergence behavior, the connection relationships are screened. The connection relationships that are retained during the structural domestication process are extracted, while the connection relationships that are stripped are excluded, thus generating the domestication and evolution structure after the connection relationship is screened. Based on the domestication and evolution structure after filtering the connection relationships, the connection relationships are collected along the connection paths, and the retained connection relationships are aggregated according to the connection paths to generate the domestication and evolution structure after the connection relationships are collected. Based on the domestication and evolution structure after the connection relationships are aggregated, the connection relationships are divided into states, and the connection relationships are divided into different structurally domesticable states, generating a set of structurally domesticable states.

[0011] Optionally, the generation of the financial data analysis results specifically includes: Map the structurally tamable states in the set of structurally tamable states to the taming evolution structure to generate a structural taming regulation mapping relationship; Based on the structural domestication regulation mapping relationship, the domestication evolution structure is divided, and the domestication evolution structure corresponding to different structural domestication states is distinguished to generate domestication evolution structures divided according to structural domestication states. The domestication and evolution structures classified according to their domesticability are integrated to generate financial data analysis results.

[0012] An enterprise financial data analysis system based on big data processing according to an embodiment of the present invention includes: The data acquisition and structure generation module is used to collect multi-source heterogeneous financial data from enterprises, perform partitioning and parallel structured processing, and generate a global financial relationship structure and a set of partitioned structures. The structural etching processing module is used to perform hierarchical structural etching on the global financial relationship structure, generate an etched skeleton structure, and constrain the etching process based on the capital flow path and transaction relationship. The etching evolution module is used to perform continuous structural etching evolution on the etched skeleton structure. It exposes deep connection relationships and stabilizes the structural core through progressive etching operations, generating an etched stable structure. The structure domestication module is used to inject constraint drive into the etched stable structure and generate domesticated evolution structure through constraint propagation and structure convergence evolution. The state partitioning module is used to perform structural convergence behavior analysis on the domestication and evolution structure. Based on the convergence behavior of the structure in the structural domestication process, it partitions the structure into domesticable states and generates a set of domesticable states. The results generation module is used to construct structural domestication adjustment mapping relationships based on the set of structurally domesticable states, and generate financial data analysis results.

[0013] The beneficial effects of this invention are: This invention addresses the problems of existing financial data analysis methods that rely solely on static relationship modeling, struggle to identify redundant connections, and fail to depict deep structural evolution. It proposes a big data-based enterprise financial data analysis method. By partitioning and parallel structuring multi-source heterogeneous financial data, a global financial relationship structure and a set of partitioned structures are constructed, allowing previously scattered data to be expressed within a unified structure. Furthermore, through hierarchical structural erosion and progressive erosion evolution, connections are stripped away at each level, removing redundant connections while retaining deep, key connections, thereby reducing structural complexity and highlighting core structural features. On this basis, the introduction of capital flow paths and transaction relationships constrains the structure, ensuring that the erosion process depends not only on the structure itself but also on the constraints of actual business logic, thus improving the authenticity and effectiveness of the structural representation.

[0014] After structural etching is completed, this invention further utilizes the constraint propagation of account connection relationships and the structural convergence evolution mechanism to allow the connection relationships to be transmitted along the association relationships in the structure and gradually converge to form a stable and consistent domestication evolution structure. Subsequently, by analyzing the retention and stripping situations during the structural domestication process, a set of domesticable structural states is generated, and a structural domestication adjustment mapping relationship is constructed based on this set of states to reflect abstract state information into specific structures, achieving fine division and reconstruction of the structure. Finally, by integrating the divided structures, financial data analysis results are generated. Compared with existing technologies, this invention can effectively strip redundant information and strengthen key connection relationships in complex financial relationship networks, and achieve multi-level analysis by combining business data constraints and structural evolution processes, thereby significantly improving the accuracy and structural expressiveness of financial data analysis, while enhancing the interpretability and stability of the analysis results. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method and system for analyzing enterprise financial data based on big data processing, as proposed in this invention. Figure 2 This is a schematic diagram of the etched stable structure generation structure of an enterprise financial data analysis method and system based on big data processing proposed in this invention. Figure 3 This is a schematic diagram illustrating the structural adaptation and adjustment mapping and result generation of an enterprise financial data analysis method and system based on big data processing proposed in this invention. Detailed Implementation

[0016] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0017] refer to Figures 1-3 A method for analyzing enterprise financial data based on big data processing includes the following steps: Collect multi-source heterogeneous financial data from enterprises, including business transaction data, cash flow data, and account association data. Perform partitioning and parallel structured processing, perform relationship association and structure expression within each partition, and perform cross-partition fusion on the structure of each partition to generate a global financial relationship structure and a set of partition structures. A hierarchical structure etching process is performed on the global financial relationship structure. In each partition of the partition structure set, the connection relationship is weakened and redundant paths are stripped in parallel. The etching results of each partition are then aggregated globally to generate an etched skeleton structure. The etching process is constrained based on the capital flow path in the capital flow data and the transaction relationship in the business transaction data. Continuous structural etching evolution is performed on the etched framework structure. Deep connections are exposed and the structural core is stabilized through progressive etching operations, resulting in an etched stable structure. Constraint driving is injected into the etched stable structure and structural acclimatization processing is performed. The constraint driving comes from the account connection relationship in the account association data. The acclimatization evolution structure is generated through constraint propagation and structural convergence evolution. Perform structural convergence behavior analysis on the domestication and evolution structure, divide the structure into domesticable states based on the convergence behavior of the structure in the structural domestication process, and generate a set of domesticable states. Based on the set of structurally tamable states, a structural taming adjustment mapping relationship is constructed to generate financial data analysis results.

[0018] In this embodiment, the generation of the global financial relationship structure and partition structure set specifically includes: Collect multi-source heterogeneous financial data from enterprises, and perform unified identification and time alignment processing on the multi-source heterogeneous financial data to form relatable data under a unified time benchmark; The associative data is divided according to the data source attribute, business type attribute, and time interval attribute, and a set of partition structure is generated; The generation of the partition structure set specifically includes: Based on data source attributes, the first level of data partitioning is performed, grouping data from the same source into the same data partition, generating a data set partitioned by data source attribute. Within this data set partitioned by data source attribute, the second level of partitioning is performed based on business type attribute, grouping data from the same business type into the same business partition, generating a data set partitioned by business type attribute. Within this data set partitioned by business type attribute, the third level of partitioning is performed based on time interval attribute, grouping data within the same time interval into the same time partition, generating a data set partitioned by time interval attribute. The data in each time partition is then integrated to generate the corresponding partition structure, and all partition structures are aggregated to generate a partition structure set. Parallel structuring is performed within each partition of the partition structure set. Field parsing and data mapping are performed on the associative data within the partition to convert multi-source heterogeneous financial data into the corresponding structural expression form and generate the structure of each partition. The generation of each partition structure specifically includes: The process involves parsing the related data within each partition, splitting the data fields to obtain the corresponding field data; mapping the field data based on the field data, unifying the mapping of data fields from different sources into a consistent field expression form, and generating mapped field data; combining the mapped field data according to the correspondence between the data to form a structural expression form; and integrating the structural expressions to generate the structure of each partition. Within the partition structure, relationship association processing is performed, transaction relationships between data are established based on business transaction data, and fund flow paths between data are established based on fund flow data, forming corresponding relationship connections within the partition structure; The generation of relational connections specifically includes: The business transaction data within the partition structure is parsed, and the related fields in the transaction records are extracted. Data with corresponding transaction relationships are connected in the partition structure to generate transaction relationships. Based on the fund flow data, the fund flow records in the partition structure are extracted, and connections are constructed according to the order of fund flow to generate fund flow paths. The connections in the transaction relationships are aligned with the connections in the fund flow paths. Connections that exist in both transaction relationships and fund flow paths are retained, while connections that do not appear in the fund flow paths are filtered out to generate filtered connections. Based on the fund flow paths, data that do not form connections in the transaction relationships are supplemented with connections, and the supplemented connections are written into the partition structure to generate relationship connections. Introduce account-related data into each partition structure, perform connection relationship construction and expansion processing on the partition structure, embed account connection relationships into relationship connections and complete structural integration to form a partition structure with a consistent structure. The generation of a consistent partition structure specifically includes: Account association data is introduced into the partition structure. The account connection relationships in the account association data are extracted and written into the connection relationship of the partition structure to generate a partition structure containing account connection relationships. Based on the partition structure containing account connection relationships, the connection relationships are expanded by adding the existing connection relationships in the account connection relationships into the connection relationship of the partition structure to generate an expanded partition structure. The connection relationships in the expanded partition structure are then sorted out, and duplicate connection relationships are merged to generate a partition structure with a consistent structure. Perform cross-partition fusion processing on the partition structure set, connect and integrate the partition structures based on the temporal continuity between partitions, generate a global financial relationship structure, and output the partition structure set synchronously. The generation of the global financial relationship structure specifically includes: The process involves extracting connections from the partition structure set and identifying connections that exist simultaneously in different partition structures. Based on these simultaneous connections, the connections in each partition structure are linked to generate an initial global financial relationship structure. Then, based on temporal continuity, the connections in the initial global financial relationship structure are arranged chronologically, and temporally adjacent connections are linked to generate a temporally continuous global financial relationship structure. Finally, the connections in the temporally continuous global financial relationship structure are integrated, and duplicate connections are merged to generate the final global financial relationship structure. The global financial relationship structure and the partition structure set are output simultaneously.

[0019] In this embodiment, the generation of the etched skeleton structure specifically includes: Obtain the global financial relationship structure and partition structure set, perform hierarchical structure etching on the global financial relationship structure, expand the connection relationship according to the hierarchical relationship and perform layered action to generate a hierarchical connection relationship structure; The generation of a hierarchical connection structure specifically includes: Based on the relationships between connections, the connections are hierarchically divided, with connections directly related to the same connection grouped into the same level, generating an initial hierarchical division result. Based on the initial hierarchical division result, the connections are expanded level by level, arranging connections at different association distances in hierarchical order, generating a hierarchically arranged connection relationship. The hierarchically arranged connection relationships are then integrated, with the connections in each level being aggregated to generate a layered connection relationship structure. Within each partition of the partition structure set, structural etching is carried out synchronously based on the hierarchical connection relationship structure. The connection relationships in each partition are filtered, and the connection relationships that appear repeatedly at the same level are regarded as redundant connection relationships. The redundant connection relationships are stripped to generate a weakened connection relationship structure. The weakening of connection structure generation specifically includes: The connection relationships in the hierarchical connection structure are extracted, and each connection relationship is divided according to the hierarchical relationship to generate a hierarchical connection relationship. Based on the hierarchical connection relationship, the connection relationships in the same level are compared, and the connection relationships that appear repeatedly in the same level are identified as redundant connection relationships. The redundant connection relationships are stripped and removed from the connection relationships in the partition structure to generate a weakened connection relationship structure. To weaken the connection structure, the redundant connection relationships are stripped along the connection path, and the connection relationships that maintain the connectivity of the partition structure are retained, thus generating the stripped partition structure. Based on the weakened connection structure and the partitioned structure after stripping, the fund flow path in the fund flow data is introduced to constrain the changes in the connection relationship. At the same time, the transaction relationship in the business transaction data is combined to limit the scope of the connection relationship to be retained, and a constrained partitioned structure is generated. The generation of the constrained partition structure specifically includes: Based on weakening the connection structure, the connections in the partition structure are extracted and matched with the capital flow paths in the capital flow data. Connections that do not conform to the capital flow paths are stripped to generate the first connection. Based on the first connection, the connections are filtered according to the transaction associations in the business transaction data. Connections that exist in the transaction associations are retained, while connections that do not appear in the transaction associations are stripped to generate the second connection. The second connection is written into the connection structure of the partition structure, and the connection structure of the partition structure is updated to generate the constrained partition structure. The constrained partition structures within each partition are aggregated to form the corresponding partition etching results. The partitioned etching results are integrated as a whole, and the global aggregation is completed by restoring the connection relationship between the partitions to generate the etched skeleton structure. The generation of the etched framework structure specifically includes: The process involves traversing the partition structures in the partition etching results and extracting the connections between them; identifying the connections that exist simultaneously in different partition structures; connecting the connections between the partition structures based on the simultaneous connections and associating them; integrating the connections between the associating partition structures and merging duplicate connections to generate the etched skeleton structure.

[0020] In this embodiment, the generation of the etched stable structure specifically includes: The connection relationships in the etched skeleton structure are etched layer by layer, and the connection relationships are screened out layer by layer along the connection path to generate the etched skeleton structure after layer-by-layer etching. The generation of the etching framework structure after layered etching specifically includes: Based on the relationships between connections, the connections are divided into hierarchical levels to generate hierarchically divided connections. Then, based on these hierarchically divided connections, redundant connections are removed from the etched skeleton structure at each level. Finally, the connections in the etched skeleton structure after the hierarchical removal are integrated to generate a layered etched skeleton structure. In the etched skeleton structure after layered etching, the connection relationship is progressively etched, and the connection relationship is peeled off step by step according to the hierarchical order to generate the etched skeleton structure after progressive etching. The generation of the etching framework structure after progressive etching specifically includes: The connection relationships in the etched skeleton structure after layered etching are extracted and divided according to the hierarchical relationship. Based on the connection relationships divided according to the hierarchical relationship, the connection relationships are stripped off level by level in hierarchical order. In each level, the connection relationships that are redundant are removed from the connection relationships in the etched skeleton structure. The etched skeleton structure after the stripping of each level is integrated to generate the etched skeleton structure after progressive etching. In the etched skeleton structure after progressive etching, the connection relationships are screened, and the connection relationships retained after etching are extracted to generate deep connection relationships; The etching skeleton structure is connected and fixed based on deep connectivity relationships. The deep connectivity relationships are written into the etching skeleton structure to generate a stable connectivity structure. The generation of stable connection structures specifically includes: Deep connectivity relationships are extracted and retained. Based on these deep connectivity relationships, the connectivity relationships in the etched skeleton structure are replaced with deep connectivity relationships. The replaced connectivity relationships in the etched skeleton structure are then integrated, and duplicate connectivity relationships are merged to generate a stable connectivity structure. The stable connection structure is integrated as a whole to generate an etched stable structure.

[0021] In this embodiment, the generation of domestication and evolution structures specifically includes: Write the account connection relationships in the account association data into the connection relationships in the etched stable structure to form an etched stable structure containing the account connection relationships; Based on the etched stable structure containing account connection relationships, constraint propagation is performed on the connection relationships. The account connection relationships are passed level by level along the connection path to generate the etched stable structure after constraint propagation. The generation of etch-stabilized structures after constraint propagation specifically includes: The connection relationships in the etched stable structure containing account connections are extracted and used as initial constraint connection relationships. Based on the initial constraint connection relationships, the connection relationships are propagated, with connection relationships associated with account connections used as propagation objects, and the account connection relationships are written into the connection relationships corresponding to the propagation objects to generate extended connection relationships. Based on the extended connection relationships, the connection relationships are propagated level by level according to the association between the connection relationships, and the account connection relationships are written into the corresponding connection relationships at each level of propagation. The connection relationships after the level-by-level propagation are integrated to generate the etched stable structure after constraint propagation. In the etched stable structure after constraint propagation, the connection relationships are filtered, retaining those consistent with the account connection relationships and stripping those that do not conform to the account connection relationships, thus generating the etched stable structure after constraint filtering. Based on the etch-stable structure after constraint screening, the connection relationships are subjected to structural convergence evolution. The retained connection relationships are then aggregated along the connection path to generate the converged etch-stable structure. The generation of the etch-stable structure after convergence specifically includes: The connectivity relationships in the etched stable structure after constraint screening are extracted, and the connectivity relationships retained during the constraint screening process are retained connectivity relationships. Based on the retained connectivity relationships, the connectivity relationships are connected, and the connectivity relationships that are related to the retained connectivity relationships are connected, so that the connectivity relationships form an associated structure in the etched stable structure. The connectivity relationships that form the associated structure are integrated, and the interconnected connectivity relationships are merged to generate the converged etched stable structure. The converged etched stable structure is integrated as a whole to generate the domestication and evolution structure.

[0022] In this embodiment, the generation of the set of structurally tunable states specifically includes: The connection relationships in the domestication and evolution structure are expanded along the connection paths, and the connection paths and node positions of each connection relationship in the structure are recorded to generate the domestication and evolution structure after the connection paths are expanded. Based on the domestication evolution structure after the connection path is expanded, the structural convergence behavior of the connection relationship is analyzed. The retention and stripping of the connection relationship in the structural domestication process are identified along the connection path, and the domestication evolution structure corresponding to the structural convergence behavior is generated. The generation of domestication and evolution structures corresponding to structural convergence behavior specifically includes: The connection relationships in the domestication evolution structure after the connection path is expanded are extracted. The connection relationships that are retained during the structural domestication process are taken as retained connection relationships, and the connection relationships that are stripped are taken as stripped connection relationships. Based on the retained connection relationships and stripped connection relationships, the connection relationships in the domestication evolution structure after the connection path is expanded are mapped to correspond to each connection relationship with its retention or stripping status during the structural domestication process, thereby generating the domestication evolution structure corresponding to the structural convergence behavior. In the domestication and evolution structure corresponding to the structural convergence behavior, the connection relationships are screened. The connection relationships that are retained during the structural domestication process are extracted, while the connection relationships that are stripped are excluded, thus generating the domestication and evolution structure after the connection relationship is screened. Based on the domestication and evolution structure after filtering the connection relationships, the connection relationships are collected along the connection paths, and the retained connection relationships are aggregated according to the connection paths to generate the domestication and evolution structure after the connection relationships are collected. Based on the domestication and evolution structure after the connection relationship is collected, the connection relationship is divided into different structurally domesticable states, generating a set of structurally domesticable states. State division specifically includes: The connection relationships in the domestication evolution structure after the connection relationships are collected are extracted to obtain the retention or separation status of each connection relationship during the structural domestication process; based on the retention or separation status, the connection relationships are divided, and connection relationships with the same retention or separation status are divided into the same structural domestication state; the connection relationships in each structural domestication state are aggregated to generate a set of structural domestication states.

[0023] In this embodiment, the generation of financial data analysis results specifically includes: Map the structurally tamable states in the set of structurally tamable states to the taming evolution structure to generate a structural taming regulation mapping relationship; The generation of structural domestication regulation mapping relationships specifically includes: The process involves extracting each tradable state from the set of tradable structural states and obtaining the connectivity relationships contained in each state; extracting connectivity relationships from the domestication evolution structure; matching these connectivity relationships within the domestication evolution structure based on the connectivity relationships contained in each tradable state, and using the matched connectivity relationships as corresponding connectivity relationships; associating each tradable structural state with its corresponding connectivity relationship, and aggregating the corresponding connectivity relationships belonging to the same tradable structural state to form a mapping relationship between tradable structural states and their corresponding connectivity relationships; and integrating these mapping relationships to generate a structural domestication regulation mapping relationship. Based on the structural domestication regulation mapping relationship, the domestication evolution structure is divided, and the domestication evolution structure corresponding to different structural domestication states is distinguished to generate domestication evolution structures divided according to structural domestication states. The generation of domestication evolution structures, classified according to their domesticability state, specifically includes: Based on the structural domestication regulation mapping relationship, the connection relationships in the domestication evolution structure are extracted; according to the structural domestication regulation mapping relationship, the connection relationships in the domestication evolution structure are assigned to the corresponding structural domestication states, and the connection relationships belonging to the same structural domestication state are classified; the connection relationships corresponding to each structural domestication state are collected separately, and the connection relationships corresponding to the same structural domestication state constitute the corresponding domestication evolution structure; the domestication evolution structure corresponding to each structural domestication state is output to generate the domestication evolution structure divided by structural domestication state. The domestication and evolution structures classified according to their domesticability are integrated to generate financial data analysis results. Integrating the domestication and evolutionary structure specifically includes: The connection relationships in the domestication evolution structure divided according to the domestication state are extracted; based on the connection relationships corresponding to each domestication state, the connection relationships are aggregated to generate a set of connection relationships corresponding to each domestication state; based on the set of connection relationships corresponding to each domestication state, the distribution of the domestication state in the domestication evolution structure is integrated to generate financial data analysis results.

[0024] A big data-based enterprise financial data analysis system includes: The data acquisition and structure generation module is used to collect multi-source heterogeneous financial data from enterprises, perform partitioning and parallel structured processing, and generate a global financial relationship structure and a set of partitioned structures. The structural etching processing module is used to perform hierarchical structural etching on the global financial relationship structure, generate an etched skeleton structure, and constrain the etching process based on the capital flow path and transaction relationship. The etching evolution module is used to perform continuous structural etching evolution on the etched skeleton structure. It exposes deep connection relationships and stabilizes the structural core through progressive etching operations, generating an etched stable structure. The structure domestication module is used to inject constraint drive into the etched stable structure and generate domesticated evolution structure through constraint propagation and structure convergence evolution. The state partitioning module is used to perform structural convergence behavior analysis on the domestication and evolution structure. Based on the convergence behavior of the structure in the structural domestication process, it partitions the structure into domesticable states and generates a set of domesticable states. The results generation module is used to construct structural domestication adjustment mapping relationships based on the set of structurally domesticable states, and generate financial data analysis results.

[0025] Example 1: To verify the feasibility of this invention in practice, it was applied to the financial data processing scenario of a large manufacturing enterprise. This enterprise has developed a financial data system covering multiple business processes, including procurement, production, sales, and fund settlement, over its long-term operations. Its data sources include business transaction data, fund flow data, and account-related data. Due to the complexity of the business chain, there are numerous overlapping relationships between various types of data, resulting in a densely connected structure, numerous redundant paths, and unclear structural hierarchy in the overall financial structure. Under traditional processing methods, relevant personnel often analyze the data through simple summary statistics or single-relationship modeling, making it difficult to identify truly crucial connections and understand the deep connections between fund flow and business behavior from an overall structural perspective, thus affecting the accuracy and reliability of the analysis results.

[0026] In this scenario, the first step is to uniformly collect multi-source, heterogeneous financial data generated by the enterprise across different business stages. The data is then partitioned and processed in parallel with structural methods. By extracting transaction relationships from business transaction data and combining this with the capital flow paths from cash flow data, the originally scattered data is transformed into a partitioned structure with a structured representation. During this process, data from different time periods and business sources are divided into different partitions, and corresponding connections are established within each partition, ensuring traceability and relevance of the data within the structure. Subsequently, through cross-partition fusion processing, the temporally continuous connections between the partitions are linked and integrated to form a unified global financial relationship structure, thus providing the foundational structure for subsequent analysis.

[0027] After the structure is built, a hierarchical etch process is performed on the global financial relationship structure. By expanding the connections hierarchically, the connections in the structure gradually emerge according to the hierarchical relationship. Within each partition, connections at the same level are filtered, and duplicate connections are stripped as redundant connections. In this process, redundant connections are gradually removed, and only the key connections that can maintain the connectivity of the structure are retained, thereby effectively reducing the complexity of the structure. At the same time, the fund flow path and transaction relationship are introduced to constrain the etch process, so that the scope of retained connections always conforms to the actual business logic, avoiding the accidental removal of important relationships due to simple structural processing. By integrating the etch results of each partition, an etched skeleton structure is formed, making the structure clearer and more hierarchical.

[0028] Based on the etched skeleton structure, a continuous structural etching evolution is further performed. By peeling off the connections layer by layer in a hierarchical order, redundant connections that recur in the structure are removed layer by layer, while key connections are retained during each etching process. This allows the structure to gradually shrink towards the core. After progressive etching, the connections that have not been peeled off are extracted to form deep connections, which are then rewritten into the etched skeleton structure, thereby generating a stable connection structure. This process transforms the structure from its original complex state to a stable state dominated by key connections.

[0029] After obtaining a stable connection structure, account connections from the account association data are introduced. Constraint-driven processing is then applied to the structure, with account connections being written into it and propagated hierarchically along the relationships, continuously expanding the constraints within the structure. Subsequently, a filtering operation retains connections consistent with the account connections while removing connections that do not meet the constraints, further unifying the structure. Based on this, through structural convergence evolution, the retained connections are aggregated and integrated, gradually forming a stable structural distribution from the originally dispersed connections, ultimately generating the domestication evolution structure.

[0030] After the domestication and evolution structure is formed, structural convergence behavior analysis is performed on the connections within the structure. By extracting the retention and stripping of connections during the structural domestication process, the connections are divided into different state categories, generating a set of structurally domesticable states. Subsequently, each state in the set of structurally domesticable states is mapped to a connection in the domestication and evolution structure. Through matching and association, a structural domestication regulation mapping relationship is established. Based on this, the connections in the domestication and evolution structure are allocated and classified according to the mapping relationship, so that the connections of different states form corresponding structural regions. Finally, the financial data analysis results are generated by integrating the various structural regions.

[0031] Table 1. Comparative Evaluation of Financial Data Analysis Methods Based on Structural Etching and Domestication Mechanisms

[0032] As shown in Table 1, in terms of data processing time, the method of this invention reduces the time from 126 minutes for the traditional rule analysis method to 72 minutes, and also shows a significant decrease compared to the 104 minutes for the relational network analysis method. This result indicates that the present invention, through "partitioned parallel structured processing and hierarchical structure etching," effectively reduces the computational redundancy caused by direct processing on the global structure, making the data processing path clearer and thus shortening the overall processing flow. In the redundant connection ratio index, the traditional method is 42.5%, the relational network method is 36.8%, while the present invention reduces it to 21.4%. This result shows that the present invention, by identifying and progressively stripping redundant connections at the same level, and by combining constraints with capital flow paths and transaction correlations, continuously reduces redundant connections during structural evolution, thereby significantly reducing structural noise.

[0033] In terms of key connection identification rate, this invention achieves 82.1%, significantly higher than the 63.2% of traditional methods and 68.7% of relationship network methods. This improvement mainly stems from the mechanism of "progressive etching and deep connection extraction," which allows connections that have not been stripped to be retained in multiple rounds of structural screening, thus reflecting core business relationships more centrally. Regarding structural stability fluctuation and structural expression consistency deviation, this invention achieves 7.9% and 9.8% respectively, both significantly lower than the comparative methods. This indicates that through "constraint propagation and structural convergence evolution," the distribution of connections in the structure is more stable, reducing overall structural fluctuations caused by changes in local relationships, thereby resulting in a more consistent structural expression.

[0034] Regarding the reconstruction error rate of the analysis results, this invention achieves 8.6%, a significant decrease compared to the 19.7% of traditional methods and 15.2% of relational network methods. This result indicates that this invention, through a "structurally manageable state partitioning and state mapping mechanism," reflects the behavioral results during the structural evolution process into the final analysis structure, making the results more closely match the true relationships of the original data. In terms of multi-source data fusion bias, this invention controls it at 7.4%, significantly lower than the comparative methods. This is because the introduction of dual constraints from business transaction data and capital flow data during the structural etching process ensures that the fusion process no longer relies solely on structural similarity but also simultaneously satisfies business logic consistency. Finally, in the effective retention rate of connection relationships, this invention reaches 78.6%, significantly higher than the 58.1% of traditional methods. This indicator demonstrates that this invention not only removes redundant connection relationships but also ensures that key connection relationships are continuously retained through structural fixation and convergence mechanisms, thereby improving the effective information density of the structure.

[0035] Based on the above data, it can be seen that the present invention outperforms the prior art in many aspects, such as data processing efficiency, redundancy reduction, key relationship identification, and structural stability. Its performance improvement mainly comes from the structural etching mechanism's step-by-step stripping of redundant connections, the constraint propagation mechanism's strengthening of structural consistency, and the state mapping mechanism's fine reconstruction of structural analysis results, thereby achieving efficient expression and accurate analysis of complex financial relationship networks.

[0036] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for analyzing enterprise financial data based on big data processing, characterized in that, Includes the following steps: Collect multi-source heterogeneous financial data from enterprises, including business transaction data, cash flow data, and account association data. Perform partitioning and parallel structured processing, perform relationship association and structure expression within each partition, and perform cross-partition fusion on the structure of each partition to generate a global financial relationship structure and a set of partition structures. A hierarchical structure etching process is performed on the global financial relationship structure. In each partition of the partition structure set, the connection relationship is weakened and redundant paths are stripped in parallel. The etching results of each partition are then aggregated globally to generate an etched skeleton structure. The etching process is constrained based on the capital flow path in the capital flow data and the transaction relationship in the business transaction data. Continuous structural etching evolution is performed on the etched framework structure. Deep connections are exposed and the structural core is stabilized through progressive etching operations, resulting in an etched stable structure. Constraint driving is injected into the etched stable structure and structural acclimatization processing is performed. The constraint driving comes from the account connection relationship in the account association data. The acclimatization evolution structure is generated through constraint propagation and structural convergence evolution. Perform structural convergence behavior analysis on the domestication and evolution structure, divide the structure into domesticable states based on the convergence behavior of the structure in the structural domestication process, and generate a set of domesticable states. Based on the set of structurally tamable states, a structural taming adjustment mapping relationship is constructed to generate financial data analysis results.

2. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the global financial relationship structure and partition structure set specifically includes: Collect multi-source heterogeneous financial data from enterprises, and perform unified identification and time alignment processing on the multi-source heterogeneous financial data to form relatable data under a unified time benchmark; Based on the data source attribute, business type attribute, and time interval attribute, the associative data is partitioned to generate a set of partition structures; Parallel structuring is performed within each partition of the partition structure set. Field parsing and data mapping are performed on the associative data within the partition, which are then converted into the corresponding structural expression form to generate the structure of each partition. Within the partition structure, relationship association processing is performed, transaction relationships between data are established based on business transaction data, and fund flow paths between data are established based on fund flow data, forming corresponding relationship connections within the partition structure; Introduce account-related data into each partition structure, perform connection relationship construction and expansion processing on the partition structure, embed account connection relationships into relationship connections and complete structural integration to form a partition structure with a consistent structure. Perform cross-partition fusion processing on the partition structure set, connect and integrate the partition structures based on the temporal continuity between partitions, generate a global financial relationship structure, and output the partition structure set synchronously.

3. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the etched framework structure specifically includes: Obtain the global financial relationship structure and partition structure set, perform hierarchical structure etching on the global financial relationship structure, expand the connection relationship according to the hierarchical relationship and perform layered action to generate a hierarchical connection relationship structure; Within each partition of the partition structure set, structural etching is carried out synchronously based on the hierarchical connection relationship structure. The connection relationships in each partition are filtered, and the connection relationships that appear repeatedly at the same level are regarded as redundant connection relationships. The redundant connection relationships are stripped to generate a weakened connection relationship structure. To weaken the connection structure, the redundant connection relationships are stripped along the connection path, and the connection relationships that maintain the connectivity of the partition structure are retained, thus generating the stripped partition structure. Based on the weakened connection structure and the partitioned structure after stripping, the fund flow path in the fund flow data is introduced to constrain the changes in the connection relationship. At the same time, the transaction relationship in the business transaction data is combined to limit the scope of the connection relationship to be retained, and a constrained partitioned structure is generated. The constrained partition structures within each partition are aggregated to form the corresponding partition etching results. The partitioned etching results are integrated as a whole, and the global aggregation is completed by restoring the connection relationship between the partitions to generate the etched skeleton structure.

4. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the etch-stabilized structure specifically includes: The connection relationships in the etched skeleton structure are etched layer by layer, and the connection relationships are screened out layer by layer along the connection path to generate the etched skeleton structure after layer-by-layer etching. In the etched skeleton structure after layered etching, the connection relationship is progressively etched, and the connection relationship is peeled off step by step according to the hierarchical order to generate the etched skeleton structure after progressive etching. In the etched skeleton structure after progressive etching, the connection relationships are screened, and the connection relationships retained after etching are extracted to generate deep connection relationships; The etching skeleton structure is connected and fixed based on deep connectivity relationships. The deep connectivity relationships are written into the etching skeleton structure to generate a stable connectivity structure. The stable connection structure is integrated as a whole to generate an etched stable structure.

5. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the domestication and evolution structure specifically includes: Write the account connection relationships in the account association data into the connection relationships in the etched stable structure to form an etched stable structure containing the account connection relationships; Based on the etched stable structure containing account connection relationships, constraint propagation is performed on the connection relationships. The account connection relationships are passed level by level along the connection path to generate the etched stable structure after constraint propagation. In the etched stable structure after constraint propagation, the connection relationships are filtered, retaining those consistent with the account connection relationships and stripping those that do not conform to the account connection relationships, thus generating the etched stable structure after constraint filtering. Based on the etch-stable structure after constraint screening, the connection relationships are subjected to structural convergence evolution. The retained connection relationships are then aggregated along the connection path to generate the converged etch-stable structure. The converged etched stable structure is integrated as a whole to generate a domestication and evolution structure.

6. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the set of structurally trainable states specifically includes: The connection relationships in the domestication and evolution structure are expanded along the connection paths, and the connection paths and node positions of each connection relationship in the structure are recorded to generate the domestication and evolution structure after the connection paths are expanded. Based on the domestication evolution structure after the connection path is expanded, the structural convergence behavior of the connection relationship is analyzed. The retention and stripping of the connection relationship in the structural domestication process are identified along the connection path, and the domestication evolution structure corresponding to the structural convergence behavior is generated. In the domestication and evolution structure corresponding to the structural convergence behavior, the connection relationships are screened. The connection relationships that are retained during the structural domestication process are extracted, while the connection relationships that are stripped are excluded, thus generating the domestication and evolution structure after the connection relationship is screened. Based on the domestication and evolution structure after filtering the connection relationships, the connection relationships are collected along the connection paths, and the retained connection relationships are aggregated according to the connection paths to generate the domestication and evolution structure after the connection relationships are collected. Based on the domestication and evolution structure after the connection relationships are aggregated, the connection relationships are divided into states, and the connection relationships are divided into different structurally domesticable states, generating a set of structurally domesticable states.

7. The enterprise financial data analysis method based on big data processing according to claim 1, characterized in that, The generation of the financial data analysis results specifically includes: Map the structurally tamable states in the set of structurally tamable states to the taming evolution structure to generate a structural taming regulation mapping relationship; Based on the structural domestication regulation mapping relationship, the domestication evolution structure is divided, and the domestication evolution structure corresponding to different structural domestication states is distinguished to generate domestication evolution structures divided according to structural domestication states. The domestication and evolution structures classified according to their domesticability are integrated to generate financial data analysis results.

8. A corporate financial data analysis system based on big data processing, executing the corporate financial data analysis method based on big data processing as described in any one of claims 1 to 7, characterized in that, include: The data acquisition and structure generation module is used to collect multi-source heterogeneous financial data from enterprises, perform partitioning and parallel structured processing, and generate a global financial relationship structure and a set of partitioned structures. The structural etching processing module is used to perform hierarchical structural etching on the global financial relationship structure, generate an etched skeleton structure, and constrain the etching process based on the capital flow path and transaction relationship. The etching evolution module is used to perform continuous structural etching evolution on the etched skeleton structure. It exposes deep connection relationships and stabilizes the structural core through progressive etching operations, generating an etched stable structure. The structure domestication module is used to inject constraint drive into the etched stable structure and generate domesticated evolution structure through constraint propagation and structure convergence evolution. The state partitioning module is used to perform structural convergence behavior analysis on the domestication and evolution structure. Based on the convergence behavior of the structure in the structural domestication process, it partitions the structure into domesticable states and generates a set of domesticable states. The results generation module is used to construct structural domestication adjustment mapping relationships based on the set of structurally domesticable states, and generate financial data analysis results.

Images