Data fusion system and method based on logistics system and financial system

Abstract

The invention discloses a data fusion system and method based on a logistics system and a financial system, and relates to the technical field of data fusion, and the system comprises a multi-source data collection layer which is used for capturing the logistics state in real time through an Internet of Things device, connecting the financial system to obtain main data of commodities and clients, and filtering abnormal data through a data cleaning rule engine; the business and financial mapping layer is used for dynamically matching logistics events and accounting subjects through a weighted Bayesian model based on a preset rule base to realize intelligent mapping; the automatic processing layer is used for generating an accounting document in real time according to a mapping result and detecting an abnormal pipeline by adopting a convolutional neural network; and the unified data layer is used for constructing a fusion warehouse containing a logistics fact table and a financial dimension table based on data output by the acquisition layer and a mapping result of the mapping layer, forming a unified data view through data fusion, and providing a cross-system BI board to support correlation analysis. According to the invention, data fusion of a logistics system and a financial system can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of data fusion, and in particular to a data fusion system and method based on a logistics system and a financial system. Background Art

[0002] Currently, enterprises generally integrate logistics management systems (such as WMS and TMS) with financial systems (such as SAP and UFIDA ERP systems) through standardized APIs. This approach typically involves two approaches: The first involves scheduled batch data transfer, where the logistics system pushes business documents to the financial system in batches on a fixed schedule (usually every night) after completing operations such as warehousing, shipping, and receipt. The second involves manually driven financial processes, where financial personnel manually match shipping documents, receipts, and sales orders, verifying information such as quantity and amount, before generating accounting vouchers in the financial system.

[0003] While this model of connecting logistics management systems (such as WMS and TMS) with financial systems (such as SAP and UFIDA ERP) through standardized APIs has initially achieved system connectivity, it suffers from the following structural flaws:

[0004] 1. Data silos lead to business disconnects. Due to the T+1 batch processing mechanism (i.e., logistics data for the day is not synchronized to the financial system until the next day), there is a significant time lag between logistics status and financial data. For example, a warehouse outbound operation completed at 12:00 noon will not be available in the financial system until the following morning; and goods signed for by a customer on the same day will not be available to the finance department until the next day. This fragmented data timeliness prevents companies from gaining real-time visibility into the "logistics-capital flow" linkage. Managers rely on lagging data for critical decision-making (such as tracking urgent order deliveries and assessing accounts receivable risk), creating operational blind spots.

[0005] 2. Manual intervention leads to errors and compliance risks. Finance personnel must perform three highly repetitive manual operations: cross-system document matching (e.g., associating a delivery order number with a sales order); data logic verification (e.g., unit price fluctuations, freight calculation discrepancies); and manual entry of voucher accounts. According to industry research (Gartner, 2023), the error rate in manual processing is as high as 5%-8%. Common issues include: inconsistencies between document and voucher amounts, triggering subsequent audit alerts; over-invoicing due to untimely synchronization of logistics status (e.g., partial returns); and mismatched accounting accounts (e.g., including shipping costs as administrative expenses). Such errors not only increase rework costs but also potentially trigger tax compliance risks.

[0006] 3. Process delays hinder operational efficiency. From the completion of logistics operations to financial accounting, there is a three-stage delay: "data transmission → manual verification → voucher generation." System-level delays: batch processing windows lead to data backlogs (average 2-3 hours); operational-level delays: finance staff must process each order individually (taking 10-15 minutes per order); exception handling delays: reconciliation discrepancies require cross-departmental communication (average delay of over 1 hour). The overall process takes 4-6 hours per order, forcing the monthly closing cycle to be extended (usually 3-5 business days), seriously slowing the company's capital turnover efficiency.

[0007] 4. Hidden costs eat into corporate profits. To maintain this model, companies need to continuously invest in three costs:

[0008] 4.1. Direct labor costs: Based on an average of 300 orders per day for a medium-sized enterprise, 3-5 full-time reconciliation personnel are required, with annual labor costs exceeding RMB 600,000.

[0009] 4.2. Error correction costs: The average annual cost of errors-induced invoice re-issuance, account adjustments, and other derivative operations is approximately RMB 150,000.

[0010] 4.3. Opportunity cost: The loss of early payment income due to process delays (calculated based on 100 million yuan in accounts receivable, the interest loss for a one-day delay is approximately 11,000 yuan).

[0011] Comprehensive estimates (IDC, 2024) show that the traditional model causes corporate logistics-financial coordination costs to reach 0.8%-1.2% of revenue, significantly weakening competitiveness.

[0012] These shortcomings create a chain reaction of negative consequences: data lag → manual dependency → reduced efficiency → high costs → inaccurate decision-making, ultimately creating a "barrel of weakness" that hinders enterprises' digital transformation. Traditional solutions are particularly difficult to meet the demands of agile operations in scenarios with high real-time requirements, such as new retail and cross-border supply chains. Summary of the Invention

[0013] In response to the needs and shortcomings of current technological development, the present invention provides a data fusion system and method based on the logistics system and the financial system, which improves the operational efficiency, financial transparency and decision support capabilities after the data fusion of the logistics system and the financial system.

[0014] In a first aspect, the present invention provides a data fusion system based on a logistics system and a financial system. The technical solutions adopted to solve the above technical problems are as follows:

[0015] A data fusion system based on a logistics system and a financial system, characterized in that it includes:

[0016] The multi-source data collection layer is used to capture logistics status in real time through IoT devices, connect to the financial system to obtain product and customer master data, and use the data cleaning rule engine to filter abnormal data;

[0017] The business and financial mapping layer is used to dynamically match logistics events and accounting items through a weighted Bayesian model based on a preset rule base to achieve intelligent mapping;

[0018] The automated processing layer is used to generate accounting vouchers in real time based on the mapping results, store the hash values ​​of key events on the blockchain, and use convolutional neural networks to detect abnormal transactions;

[0019] The unified data layer is used to build a fusion warehouse containing logistics fact tables and financial dimension tables based on the output data of the multi-source data acquisition layer and the mapping results of the business and financial mapping layer through dimensional modeling. After data fusion, a unified data view is formed, and finally a cross-system BI dashboard is provided to support correlation analysis.

[0020] Optionally, the multi-source data acquisition layer involved specifically includes:

[0021] IoT collection module, used to capture logistics status data in real time through RFID or GPS devices;

[0022] System docking module, used to synchronize with the financial system and obtain product master data and customer master data;

[0023] The data cleaning module is used to set the time interval threshold and spatial displacement threshold of the logistics trajectory, and use the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

[0024] Optionally, the business and financial mapping layer involved specifically includes:

[0025] The rule management module supports a visual rule configuration interface, allowing financial experts to customize the mapping relationship between logistics events and accounting subjects to form a preset rule library;

[0026] The Bayesian reasoning module is used to build a weighted Bayesian model. Based on the type, timestamp, and business attributes of logistics events, it calculates the probability of matching with accounting items. It also sets confidence intervals for amount thresholds or time window parameters to improve mapping accuracy.

[0027] The dynamic adjustment module is used to dynamically adjust the weight parameters of the Bayesian model based on the historical mapping accuracy to optimize the account mapping rules, and automatically identify and correct mapping deviations caused by changes in accounting standards by continuously learning historical mapping data.

[0028] Optionally, the automation processing layer involved specifically includes:

[0029] The voucher generation module is used to automatically generate four-part accounting vouchers based on the mapping results of logistics events and accounting subjects;

[0030] The blockchain evidence storage module is used to store the hash values ​​of logistics receipt events and voucher generation events on the blockchain to ensure audit traceability;

[0031] The anomaly detection module is used to build a convolutional neural network model, which performs real-time analysis of transaction data and identifies abnormal patterns such as sudden changes in amounts or incorrect account mappings.

[0032] Optionally, the unified data layer involved specifically includes:

[0033] The dimensional modeling module is used to construct logistics fact tables and financial dimension tables using a star schema based on the output data of the multi-source data acquisition layer and the mapping results of the business and financial mapping layer;

[0034] The data fusion module is used to associate and integrate logistics fact tables and financial dimension tables through common key fields, establish a cross-system data mapping relationship table, and form a unified data view that includes "logistics events-accounting subjects-time dimensions";

[0035] The BI dashboard module is used to provide a cross-system analysis dashboard based on a unified data view. By integrating the "logistics event-accounting subject-time dimension" data, it supports the "logistics-capital flow" linkage analysis function and the accounts receivable risk warning function.

[0036] Preferably, when constructing the logistics fact table, the dimensional modeling module sets multiple time dimensions including order creation time, logistics completion time, and financial confirmation time to support analysis requirements of different business cycles.

[0037] In a second aspect, the present invention provides a data fusion method based on a logistics system and a financial system. The technical solution adopted to solve the above technical problems is as follows:

[0038] A data fusion method based on a logistics system and a financial system comprises the following steps:

[0039] S1. Real-time logistics status data is captured through IoT devices, while simultaneously connecting to the financial system to obtain product and customer master data. The data cleaning rule engine is then used to filter abnormal data and complete missing fields, providing standardized data for subsequent processing.

[0040] S2. Based on the output data of step S1, dynamically match logistics events with accounting items through a preset rule base and a weighted Bayesian model to achieve intelligent mapping;

[0041] S3. Generate accounting vouchers in real time based on the mapping results, use blockchain to store hash values ​​of key events, and use convolutional neural networks to detect anomalies in the flow data.

[0042] S4. Based on the output data of step S1 and the business mapping results of step S2, a fusion warehouse containing logistics fact tables and financial dimension tables is constructed through dimensional modeling. A unified data view is formed through data fusion, and finally a cross-system BI dashboard is provided to support association analysis.

[0043] Optionally, executing step S1, capturing logistics status data in real time through RFID or GPS equipment;

[0044] The data cleaning rule engine sets the time interval threshold and spatial displacement threshold of the logistics trajectory, and uses the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

[0045] Optionally, step S2 specifically includes:

[0046] S2.1. Through a visual rule configuration interface, financial experts are allowed to customize the mapping relationship between logistics events and accounting subjects, forming a preset rule library;

[0047] S2.2. The weighted Bayesian model calculates the probability of matching with accounting items based on the type, timestamp, and business attributes of logistics events. It also sets confidence intervals for amount thresholds or time window parameters to improve mapping accuracy.

[0048] S2.3. Dynamically adjust the weight parameters of the Bayesian model based on the historical mapping accuracy to optimize the account mapping rules, and automatically identify and correct mapping deviations caused by changes in accounting standards by continuously learning historical mapping data.

[0049] Optionally, step S4 may include:

[0050] S4.1. Based on the output data and business from step S1 and the mapping results from step S2, construct the logistics fact table and financial dimension table using a star schema.

[0051] S4.2. Link and integrate the logistics fact table and financial dimension table using common key fields, establish a cross-system data mapping relationship table, and form a unified data view that includes "logistics events - accounting subjects - time dimensions";

[0052] S4.3. Provide a cross-system analysis dashboard based on a unified data view. By integrating the "logistics event-accounting subject-time dimension" data, it supports the "logistics-capital flow" linkage analysis function and the accounts receivable risk warning function.

[0053] The data fusion system and method based on the logistics system and the financial system of the present invention have the following beneficial effects compared with the prior art:

[0054] 1. This invention uses IoT devices to capture logistics status data in real time, achieving millisecond-level collection of information such as transportation tracks and receipt times, reducing data latency to less than 500 milliseconds. It connects to the financial system to obtain product and customer master data, forming a cross-system basic data pool. The data cleaning rule engine automatically filters abnormal tracks using a spatiotemporal threshold algorithm and automatically completes incomplete data such as missing logistics node times, thereby ensuring the integrity of the original data.

[0055] 2. Based on a preset rule base and a weighted Bayesian model, this invention calculates the probability of accounting account matching based on logistics event type, timestamp, and business attributes. The confidence interval setting for the amount threshold significantly improves mapping accuracy. It automatically optimizes model weights using historical mapping data, shortens the recognition and response cycle to accounting standard changes, achieves adaptive evolution of account mapping rules, and reduces manual rule maintenance costs.

[0056] 3. The present invention generates four-part accounting vouchers in real time based on the mapping results, improving processing efficiency. The blockchain stores hash values ​​of key events such as logistics receipt and voucher generation, reducing the audit traceability delay to less than 10 seconds. The convolutional neural network model analyzes transaction data in real time, improving the recognition rate of abnormalities such as sudden changes in amounts or account mapping errors, shortening the early warning response time, and significantly reducing the human error rate.

[0057] 4. Based on the cleaned collected data and mapping results, the present invention constructs a logistics fact table and a financial dimension table, and forms a unified data view containing multiple dimensions through the association of common key fields, which greatly shortens the data aggregation query response time; the cross-system BI dashboard supports the "logistics-capital flow" linkage analysis, providing real-time decision-making data support for management. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Attachment Figure 1 This is a connection block diagram of the first embodiment of the present invention;

[0059] Attachment Figure 2 This is a flow chart of the method of embodiment 2 of the present invention. DETAILED DESCRIPTION

[0060] In order to make the technical solution, the technical problems solved and the technical effects of the present invention more clear, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0061] Example 1:

[0062] Combined with attachment Figure 1 This embodiment proposes a data fusion system based on a logistics system and a financial system, which is characterized by including:

[0063] The multi-source data collection layer is used to capture logistics status in real time through IoT devices, connect to the financial system to obtain product and customer master data, and use the data cleaning rule engine to filter abnormal data;

[0064] The business and financial mapping layer is used to dynamically match logistics events and accounting items through a weighted Bayesian model based on a preset rule base to achieve intelligent mapping;

[0065] The automated processing layer is used to generate accounting vouchers in real time based on the mapping results, store the hash values ​​of key events on the blockchain, and use convolutional neural networks to detect abnormal transactions;

[0066] The unified data layer is used to build a fusion warehouse containing logistics fact tables and financial dimension tables based on the output data of the multi-source data acquisition layer and the mapping results of the business and financial mapping layer through dimensional modeling. After data fusion, a unified data view is formed, and finally a cross-system BI dashboard is provided to support correlation analysis.

[0067] In this embodiment, the multi-source data acquisition layer specifically includes:

[0068] IoT collection module, used to capture logistics status data in real time through RFID or GPS devices;

[0069] The system connection module is used to connect to the financial system synchronously to obtain product master data (such as code and cost price) and customer master data (such as credit rating and settlement method);

[0070] The data cleaning module is used to set the time interval threshold and spatial displacement threshold of the logistics trajectory, and use the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

[0071] In this embodiment, the business and financial mapping engine involved specifically includes:

[0072] A rule management module is configured to support a visual rule configuration interface, allowing financial experts to customize the mapping relationship between logistics events and accounting subjects, forming a preset rule library (e.g., "transportation completed → accounts receivable confirmed");

[0073] A Bayesian inference module is configured to build a weighted Bayesian model, calculate the matching probability with the accounting subject based on the type, timestamp and business attribute of the logistics event, and set a confidence interval for the amount threshold or time window parameter to improve the mapping accuracy;

[0074] A dynamic adjustment module is configured to dynamically adjust the weight parameter of the Bayesian model according to the historical mapping accuracy to optimize the subject mapping rule, and automatically identify and correct the mapping deviation caused by the change of accounting standards through continuous learning of historical mapping data.

[0075] In this embodiment, the automatic processing layer specifically includes:

[0076] A voucher generation module is configured to automatically generate four copies of accounting vouchers (e.g., "debit: accounts receivable, credit: main business income") based on the mapping results of logistics events and accounting subjects;

[0077] A blockchain storage module is configured to store the hash values of logistics signing events and voucher generation events on the blockchain to ensure audit traceability;

[0078] An anomaly detection module is configured to build a convolutional neural network model to analyze the flow data in real time and identify abnormal patterns such as sudden changes in amount or errors in subject mapping.

[0079] In this embodiment, the unified data layer specifically includes:

[0080] A dimension modeling module is configured to build a logistics fact table and a financial dimension table using a star model based on the output data of the multi-source data collection layer and the mapping results of the business and financial mapping engine;

[0081] A data fusion module is configured to associate and integrate the logistics fact table and the financial dimension table through common key fields, establish a cross-system data mapping relationship table, and form a unified data view containing "logistics event - accounting subject - time dimension";

[0082] A BI dashboard module is configured to provide a cross-system analysis dashboard based on the unified data view, support "logistics-fund flow" linkage analysis function, and accounts receivable risk warning function by integrating "logistics event - accounting subject - time dimension" data.

[0083] It should be noted that the dimension modeling module sets multiple time dimensions including order creation time, logistics completion time and financial confirmation time when building the logistics fact table, supporting analysis requirements of different business cycles.

[0084] It should be added that the accounting documents generated by the automated processing layer can also be used as input to the unified data layer to enrich the financial dimension of the unified data view.

[0085] Example 2:

[0086] Combined with attachment Figure 2 This embodiment proposes a data fusion method based on a logistics system and a financial system, which includes the following steps:

[0087] S1. Real-time capture of logistics status data through IoT devices, while connecting to the financial system to obtain product master data (such as code, cost price) and customer master data (such as credit rating, settlement method), and using the data cleaning rule engine to filter abnormal data and complete missing fields to provide standardized data for subsequent processing.

[0088] Specifically, logistics status data is captured in real time through RFID or GPS devices.

[0089] The data cleaning rule engine sets the time interval threshold and spatial displacement threshold of the logistics trajectory, and uses the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

[0090] S2. Based on the output data of step S1, the preset rule base and weighted Bayesian model are used to dynamically match logistics events with accounting items to achieve intelligent mapping. To achieve this process, the following operations are performed:

[0091] Specifically include:

[0092] S2.1. Through a visual rule configuration interface, financial experts are allowed to customize the mapping relationship between logistics events and accounting subjects, forming a preset rule library (e.g., "Transportation completed → Accounts receivable confirmed");

[0093] S2.2. The weighted Bayesian model calculates the probability of matching with accounting items based on the type, timestamp, and business attributes of logistics events. It also sets confidence intervals for amount thresholds or time window parameters to improve mapping accuracy.

[0094] S2.3. Dynamically adjust the weight parameters of the Bayesian model based on the historical mapping accuracy to optimize the account mapping rules, and automatically identify and correct mapping deviations caused by changes in accounting standards by continuously learning historical mapping data.

[0095] S3. Generate four-part accounting vouchers in real time based on the mapping results (e.g., receipt triggers "debit: accounts receivable, credit: main business income"), use blockchain to store the hash values ​​of key events, and use convolutional neural networks to detect flow data to detect anomalies.

[0096] During this step, the hash values ​​of the logistics receipt event and the voucher generation event are stored on the blockchain to ensure audit traceability. A convolutional neural network is used to detect abnormal patterns such as sudden changes in amounts or incorrect account mapping.

[0097] S4. Based on the output data of step S1 and the business mapping results of step S2, a fusion warehouse containing logistics fact tables and financial dimension tables is constructed through dimensional modeling. After data fusion, a unified data view is formed, and finally a cross-system BI dashboard is provided to support correlation analysis. This process specifically includes:

[0098] S4.1. Based on the output data and business from step S1 and the mapping results from step S2, construct the logistics fact table and financial dimension table using a star schema.

[0099] S4.2. Link and integrate the logistics fact table and financial dimension table using common key fields, establish a cross-system data mapping relationship table, and form a unified data view that includes "logistics events - accounting subjects - time dimensions";

[0100] S4.3. Provide a cross-system analysis dashboard based on a unified data view. By integrating the "logistics event-accounting subject-time dimension" data, it supports the "logistics-capital flow" linkage analysis function and the accounts receivable risk warning function.

[0101] It should be added that the accounting voucher generated in step S3 can also be used as input to step S4 to enrich the financial dimension of the unified data view.

[0102] In summary, the data fusion system and method based on the logistics system and the financial system of the present invention can improve the operational efficiency, financial transparency and decision support capabilities after the data fusion of the logistics system and the financial system.

[0103] The above specific examples are used to illustrate the principles and implementation methods of the present invention in detail. These examples are only used to help understand the core technical content of the present invention. Based on the above specific embodiments of the present invention, any improvements and modifications made by those skilled in the art without departing from the principles of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A data fusion system based on logistics system and financial system, characterized by: It includes: The multi-source data collection layer is used to capture logistics status in real time through IoT devices, connect to the financial system to obtain product and customer master data, and use the data cleaning rule engine to filter abnormal data; The business and financial mapping layer is used to dynamically match logistics events and accounting items through a weighted Bayesian model based on a preset rule base to achieve intelligent mapping; The automated processing layer is used to generate accounting vouchers in real time based on the mapping results, store the hash values ​​of key events on the blockchain, and use convolutional neural networks to detect abnormal transactions; The unified data layer is used to build a fusion warehouse containing logistics fact tables and financial dimension tables based on the output data of the multi-source data acquisition layer and the mapping results of the business and financial mapping layer through dimensional modeling. After data fusion, a unified data view is formed, and finally a cross-system BI dashboard is provided to support correlation analysis.

2. The data fusion system based on the logistics system and the financial system according to claim 1 is characterized in that: The multi-source data acquisition layer specifically includes: IoT collection module, used to capture logistics status data in real time through RFID or GPS devices; System docking module, used to synchronize with the financial system and obtain product master data and customer master data; The data cleaning module is used to set the time interval threshold and spatial displacement threshold of the logistics trajectory, and use the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

3. The data fusion system based on the logistics system and the financial system according to claim 1 is characterized in that: The business and financial mapping layer specifically includes: The rule management module supports a visual rule configuration interface, allowing financial experts to customize the mapping relationship between logistics events and accounting subjects to form a preset rule library; The Bayesian reasoning module is used to build a weighted Bayesian model. Based on the type, timestamp, and business attributes of logistics events, it calculates the probability of matching with accounting items. It also sets confidence intervals for amount thresholds or time window parameters to improve mapping accuracy. The dynamic adjustment module is used to dynamically adjust the weight parameters of the Bayesian model based on the historical mapping accuracy to optimize the account mapping rules, and automatically identify and correct mapping deviations caused by changes in accounting standards by continuously learning historical mapping data.

4. The data fusion system based on the logistics system and the financial system according to claim 1 is characterized in that: The automation processing layer specifically includes: The voucher generation module is used to automatically generate four-part accounting vouchers based on the mapping results of logistics events and accounting subjects; The blockchain evidence storage module is used to store the hash values ​​of logistics receipt events and voucher generation events on the blockchain to ensure audit traceability; The anomaly detection module is used to build a convolutional neural network model, which performs real-time analysis of transaction data and identifies abnormal patterns such as sudden changes in amounts or incorrect account mappings.

5. The data fusion system based on the logistics system and the financial system according to claim 1 is characterized in that: The unified data layer specifically includes: The dimensional modeling module is used to construct logistics fact tables and financial dimension tables using a star schema based on the output data of the multi-source data acquisition layer and the mapping results of the business and financial mapping layer; The data fusion module is used to associate and integrate logistics fact tables and financial dimension tables using common key fields, establish a cross-system data mapping relationship table, and form a unified data view that includes "logistics events, accounting subjects, and time dimensions"; The BI dashboard module is used to provide a cross-system analysis dashboard based on a unified data view. By integrating "logistics events-accounting subjects-time dimension" data, it supports the "logistics-capital flow" linkage analysis function and the accounts receivable risk warning function.

6. The data fusion system based on the logistics system and the financial system according to claim 5 is characterized in that: When constructing the logistics fact table, the dimensional modeling module sets multiple time dimensions including order creation time, logistics completion time, and financial confirmation time to support analysis requirements of different business cycles.

7. A data fusion method based on logistics system and financial system, characterized in that: The steps include: S1. Real-time logistics status data is captured through IoT devices, while simultaneously connecting to the financial system to obtain product and customer master data. The data cleaning rule engine is then used to filter abnormal data and complete missing fields, providing standardized data for subsequent processing. S2. Based on the output data of step S1, dynamically match logistics events with accounting items through a preset rule base and a weighted Bayesian model to achieve intelligent mapping; S3. Generate accounting vouchers in real time based on the mapping results, use blockchain to store hash values ​​of key events, and use convolutional neural networks to detect anomalies in the flow data. S4. Based on the output data of step S1 and the business mapping results of step S2, a fusion warehouse containing logistics fact tables and financial dimension tables is constructed through dimensional modeling. A unified data view is formed through data fusion, and finally a cross-system BI dashboard is provided to support association analysis.

8. The data fusion method based on the logistics system and the financial system according to claim 7 is characterized in that: Execute step S1 to capture logistics status data in real time through RFID or GPS equipment; The data cleaning rule engine sets the time interval threshold and spatial displacement threshold of the logistics trajectory, and uses the spatiotemporal threshold algorithm to automatically mark and filter abnormal trajectories that exceed the threshold. It also supports data integrity verification function, and automatically completes or issues warnings for incomplete data with missing logistics node time or no customer signature confirmation to ensure the quality of the original data.

9. The data fusion method based on the logistics system and the financial system according to claim 7 is characterized in that: The step S2 specifically includes: S2.

1. Through a visual rule configuration interface, financial experts are allowed to customize the mapping relationship between logistics events and accounting subjects, forming a preset rule library; S2.

2. The weighted Bayesian model calculates the probability of matching with accounting items based on the type, timestamp, and business attributes of logistics events. It also sets confidence intervals for amount thresholds or time window parameters to improve mapping accuracy. S2.

3. Dynamically adjust the weight parameters of the Bayesian model based on the historical mapping accuracy to optimize the account mapping rules, and automatically identify and correct mapping deviations caused by changes in accounting standards by continuously learning historical mapping data.

10. The data fusion method based on the logistics system and the financial system according to claim 7, characterized in that: The step S4 specifically includes: S4.

1. Based on the output data and business from step S1 and the mapping results from step S2, construct the logistics fact table and financial dimension table using a star schema. S4.

2. Use common key fields to link and integrate the logistics fact table and financial dimension table, establish a cross-system data mapping relationship table, and form a unified data view that includes "logistics events - accounting subjects - time dimensions"; S4.

3. Provide a cross-system analysis dashboard based on a unified data view. By integrating "logistics events-accounting subjects-time dimension" data, it supports the "logistics-capital flow" linkage analysis function and the accounts receivable risk warning function.

Images