ERP multi-source data fusion method and system for cross-border trade

Abstract

The application discloses an ERP multi-source data fusion method and system for cross-border trade, belongs to the technical field of ERP data processing, and obtains multi-dimensional data to be fused; in the process of cross-border trade data access, communication quality dynamic evaluation is performed based on communication quality parameters of equipment terminals, and link dynamic regulation and control are performed based on equipment data types and communication quality dynamic evaluation results; heterogeneous difference field evaluation is performed based on the difference degree of the multi-dimensional data, difference fields of the multi-dimensional data are obtained, and mapping alignment is performed based on semantic middleware and the difference fields of the multi-dimensional data to eliminate data semantic ambiguity between different systems; in the fusion modeling process, fusion granularity dynamic regulation and control are performed according to cross-border business requirements and the collaborative efficiency parameters of the multi-dimensional data, and dynamic data fusion is performed based on global data accuracy parameters, fusion granularity dynamic regulation and control results and communication quality dynamic evaluation results.

Description

Technical Field

[0001] This invention relates to the field of ERP data processing technology, and in particular to a method and system for ERP multi-source data fusion for cross-border trade. Background Technology

[0002] Cross-border trade involves a long and complex business chain, encompassing overseas marketing, international logistics, customs clearance, cross-border payments, and multi-country tax compliance. This results in businesses generating and relying on massive amounts of data from multiple heterogeneous data sources, including e-commerce platforms, independent websites, logistics service providers, supply chain vendors, and customs systems. Traditional ERP systems, as the core of internal management, face significant challenges in handling this dispersed, heterogeneous, and real-time-critical multi-source data, often leading to data silos, information fragmentation, and process disruptions. Therefore, efficiently integrating, cleaning, and fusing this multi-source data, and leveraging it to empower ERP systems to achieve integrated collaboration and data-driven decision-making from sales forecasting, intelligent procurement, inventory optimization to financial accounting, has become crucial for enhancing the core competitiveness of cross-border trade enterprises.

[0003] For example, Chinese invention patent CN120875809A discloses a product business data processing method, device, and medium for an ERP business scenario, including: upon triggering a business indicator prediction request, acquiring historical sales data, supply chain data, and external environment data of the target product in the ERP system; determining multiple dynamic time scales corresponding to the target product based on the supply chain data, and determining a multi-dimensional feature set corresponding to each dynamic time scale; determining linear predicted values ​​and nonlinear residual sequences of business indicators through an ARIMA model and each multi-dimensional feature set; decomposing and predicting the nonlinear residual sequence based on the external environment data to determine a reconstructed residual sequence; determining the current business indicator prediction data corresponding to each dynamic time scale based on the linear predicted values ​​of business indicators and the reconstructed residual sequence; and performing multi-scale fusion on the business indicator prediction data corresponding to each dynamic time scale to determine the target business indicator prediction data for the target product.

[0004] For example, Chinese invention patent CN120996449A discloses an intelligent decision-making system for mine production and operation based on multi-source heterogeneous data fusion, including: a data acquisition and access module; a data preprocessing module; a data fusion modeling module; an intelligent decision-making module; and a decision execution feedback module. The data acquisition unit deploys edge computing terminals, integrates multi-mode communication technology, and accesses data from underground sensors, electromechanical equipment, etc., in real time, ensuring data transmission reliability through a communication quality assessment formula. The business system interconnection unit uses semantic middleware to map heterogeneous fields from systems such as ERP and WMS to globally unified business keys, eliminating semantic ambiguity. The data preprocessing module integrates scattered data into a standardized dataset through intelligent cleaning, format conversion, and three-dimensional association.

[0005] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0006] In existing technologies, during the ERP multi-source data fusion process, the large differences in data types of equipment in different operational stages and the limited cross-border network resources may lead to poor stability of data transmission links at the equipment terminals, low adaptability of fusion granularity to business needs, resulting in insufficient reliability of data fusion results, low efficiency of cross-border business collaboration, and difficulty in balancing data value and processing efficiency. Summary of the Invention

[0007] To address the problems in existing ERP multi-source data fusion technologies, such as poor data transmission stability at device terminals, low adaptability of fusion granularity to business needs, insufficient reliability of data fusion results, low efficiency of cross-border business collaboration, and difficulty in balancing data value and processing efficiency, due to significant differences in equipment data types across different operational stages and limited cross-border network resources, this invention provides an ERP multi-source data fusion method and system for cross-border trade. The technical solution is as follows:

[0008] On the one hand, a method for merging multi-source data in ERP systems for cross-border trade is provided. This method includes: acquiring multi-dimensional data to be merged; during the cross-border trade data access process, performing dynamic communication quality assessment based on the communication quality parameters of the device terminal, and performing dynamic link control based on the device data type and the dynamic communication quality assessment results; performing heterogeneous difference field assessment based on the degree of difference of the multi-dimensional data to obtain the difference fields of the multi-dimensional data, and performing mapping and alignment based on the semantic middleware of the multi-dimensional data and the difference fields to eliminate data semantic ambiguity between different systems; during the fusion modeling process, performing dynamic control of fusion granularity based on cross-border business needs and the collaborative efficiency parameters of multi-dimensional data, and performing dynamic data fusion based on the accuracy parameters of the whole-domain data, the dynamic control results of fusion granularity, and the dynamic communication quality assessment results.

[0009] On the other hand, an ERP multi-source data fusion system for cross-border trade is provided. The system includes: a data acquisition module, a link control module, a mapping and alignment module, and a data fusion module. The data acquisition module acquires multi-dimensional data to be fused. The link control module performs dynamic communication quality assessment based on the communication quality parameters of the device terminal during cross-border trade data access, and performs dynamic link control based on the device data type and the dynamic communication quality assessment results. The mapping and alignment module performs heterogeneous difference field assessment based on the degree of difference in multi-dimensional data to obtain the difference fields of the multi-dimensional data, and performs mapping and alignment based on the semantic middleware of the multi-dimensional data and the difference fields to eliminate semantic ambiguity between different systems. The data fusion module performs dynamic control of the fusion granularity based on cross-border business needs and the collaborative efficiency parameters of multi-dimensional data during fusion modeling, and performs dynamic data fusion based on the accuracy parameters of the entire data domain, the dynamic control results of the fusion granularity, and the dynamic communication quality assessment results.

[0010] Beneficial effects

[0011] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0012] 1. This invention achieves efficient and reliable data transmission, improves the consistency of multi-source data, and achieves the optimal balance between information value and processing efficiency by performing dynamic evaluation and link dynamic adjustment based on communication quality parameters of device terminals, mapping and alignment based on heterogeneous difference fields to eliminate semantic ambiguity, and performing dynamic adjustment of fusion granularity and dynamic data fusion according to cross-border business needs and collaborative efficiency parameters.

[0013] 2. This invention establishes a multi-dimensional quantitative evaluation model for connection establishment success rate, network disconnection rate, and data accuracy. Based on the evaluation results and device type, it implements a dual control mechanism of dynamic adjustment of encoding rate and regional link switching. This enables precise adaptation of bandwidth resources at the device level to optimize transmission efficiency and ensures the stability and connectivity of key data links at the regional level. In turn, it provides a reliable, efficient, and adaptable data transmission foundation for multi-source data fusion of upper-layer ERP systems in complex cross-border network environments.

[0014] 3. This invention sets the initial fusion granularity based on the type of business needs and dynamically weights and corrects it by combining the real-time deviation of key business performance indicators such as inventory, orders, and delivery. This allows the precision of data fusion to respond sensitively to the actual business operation status and efficiency fluctuations, thereby achieving an adaptive balance between the depth of information value and system processing efficiency. This ensures that the data view output by the ERP system is always highly consistent with the current focus and optimization direction of cross-border trade management.

[0015] 4. This invention maps the communication quality level of the device terminal to differentiated fusion weights and introduces an abnormal data early warning and elimination mechanism. At the same time, it dynamically allocates the remaining weights based on the accuracy of the whole domain data. This enables the final data fusion to adaptively prioritize the adoption of highly reliable and high-quality data sources and balance the contribution ratio of various types of data. In this way, it can maximize the absorption of effective information and output an authoritative data view that matches the real-time network status and data quality and has high business adaptability while ensuring the core reliability of the fusion result. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of a multi-source data fusion method for ERP systems oriented towards cross-border trade, provided in an embodiment of this application;

[0018] Figure 2 A dynamic data fusion flowchart for the ERP multi-source data fusion method for cross-border trade provided in the embodiments of this application;

[0019] Figure 3 This is a schematic diagram of the structure of an ERP multi-source data fusion system for cross-border trade provided in an embodiment of this application. Detailed Implementation

[0020] The technical solution provided in this application will now be described with reference to the accompanying drawings.

[0021] To facilitate understanding of the embodiments of this application, the following points will be explained first:

[0022] First, in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship; the specific meaning can be understood in context. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.

[0023] Second, the use of prefixes such as "first" and "second" in this application is solely for the purpose of distinguishing and describing different things belonging to the same category, and does not constrain the order, size, or quantity of things. For example, "first message" and "second message" are simply different messages, and there is no chronological, size, or priority relationship between them.

[0024] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0025] like Figure 1The diagram shows a flowchart of a multi-source data fusion method for ERP systems oriented towards cross-border trade provided in this application embodiment. The method includes the following steps: acquiring multi-dimensional data to be fused, including ERP business data, external related data, and equipment terminal data. ERP business data includes historical sales data, cross-border supply chain data, and cross-border inventory management data, used to support order, warehousing, and logistics collaborative management throughout the entire cross-border trade process. External related data includes cross-border trade market external environment data and cross-border trade policy data. Market external environment data covers market supply and demand and exchange rate fluctuation data of the target trading country, while policy data covers import and export tariff policies, trade barrier policies, and cross-border customs clearance supervision policies. Equipment terminal data related to cross-border trade includes downhole sensor data and electromechanical equipment operation data in cross-border trade operation scenarios, used to collect on-site operational status data for cross-border trade warehousing, loading and unloading, and transportation. During the access process, dynamic communication quality assessment is performed based on the communication quality parameters of the device terminal. Link dynamic adjustment is then performed based on the device data type and the results of the dynamic communication quality assessment. This link dynamic adjustment involves dynamically adjusting the link coding rate of the device terminal and performing local area link switching in the work area to optimize the reliability and efficiency of data transmission, ensuring that communication quality adapts to changes in network conditions. Heterogeneous difference field assessment is performed based on the degree of difference in multi-dimensional data, resulting in heterogeneous difference fields. This assessment includes semantic difference assessment, format difference assessment, and value range difference assessment. In semantic difference assessment, the business meaning and context of the fields are compared using an ontology or business terminology table to calculate semantic similarity. In format difference assessment, the data type, length, precision, and encoding format of the fields are directly compared to determine their technical compatibility. In value range difference assessment, the value range, enumeration list, or unit of field instances are statistically analyzed to quantify their numerical consistency. Each evaluation generates a quantified difference score for each field. When the score exceeds the preset threshold for the corresponding dimension, the field is identified as a difference field in that dimension. Finally, all marked fields across all dimensions are aggregated, deduplicated, and a complete list of difference fields is formed, serving as a clear target for subsequent mapping and alignment. Based on the semantic middleware of multi-dimensional data and the difference fields, mapping and alignment are performed to eliminate semantic ambiguity between different systems, ensuring the reliability of data transmission and the integrity of access. The semantic middleware utilizes its built-in ontology library, business rule library, and predefined mapping logic to perform parsing, transformation, and unified standardization processing on each field in the difference field list.First, semantic parsing and mapping are performed: Based on field names and context, the middleware matches core concepts from the ontology library, such as mapping the source system's "ItemID" and the target system's "ProductCode" to the core concept "unique product identifier," and recording this in the mapping rule library. Then, format standardization is performed: For format differences, the middleware calls pre-built conversion functions for automatic cleaning and conversion. These conversion functions are built into the semantic middleware and are a type of standardized processing subroutines based on cross-border trade business rules and data standards. They cover three core types: data type conversion functions, format adaptation functions, and unit conversion functions. Specifically, this includes converting character dates to ISO8601. The middleware includes standard date conversion functions, encoding conversion functions to convert different number systems to decimal encoding, unit conversion functions to convert non-standard units of measurement such as feet and pounds to international standard units such as kilograms and meters, and encoding adaptation functions to convert different character encodings (such as UTF-8 and GBK) to a unified encoding format. Its core function is to perform automated and standardized cleaning and conversion for heterogeneous data format differences, achieving unified alignment of multi-source data formats without manual intervention. Finally, it performs value domain unification: for value domain differences, the middleware performs value conversion according to business rules, such as mapping "Male / Female" to "1 / 0", or unifying different currency amounts to the base currency based on the exchange rate. Ultimately, a data model is generated in which all fields follow a unified standard in terms of meaning, format, and value. This fundamentally eliminates semantic ambiguity between different systems and lays a solid foundation for subsequent reliable data fusion and business analysis. During the fusion modeling process, dynamic adjustment of fusion granularity is performed based on cross-border business needs and collaborative efficiency parameters. Dynamic data fusion is also performed based on global data accuracy parameters, dynamic adjustment results of fusion granularity, and dynamic evaluation results of communication quality. Dynamic adjustment of fusion granularity means dynamically adjusting the granularity of data fusion to achieve the optimal balance between information value and processing efficiency. Dynamic data fusion means dynamically adjusting the fusion weights of multi-dimensional data to ensure the reliability and business adaptability of the fusion results.

[0026] In this embodiment, the present invention achieves efficient and reliable fusion of multi-source ERP data for cross-border trade through a full-process design that includes precise multi-dimensional data access, standardized alignment of heterogeneous difference fields, dynamic link control, and fusion granularity adaptation. By classifying data sources into three categories—ERP business data, external related data, and device terminal data—and clarifying data boundaries and application scenarios, and combining dynamic evaluation of device communication quality parameters with link coding rate control and local area link switching, the invention effectively solves the problems of high latency and high packet loss rate in device terminal data transmission under fluctuating cross-border network environments, ensuring the real-time performance and integrity of multi-source data access. By constructing a three-layer heterogeneous difference field evaluation system based on semantics, format, and value domain, the invention accurately identifies heterogeneous conflict points in multi-source data and, relying on semantic middleware based on the built-in ontology library and business rule library, completes the unambiguous mapping of core concepts of fields. Pre-defined conversion functions drive format standardization and value domain unification under business rule constraints, completely eliminating semantic ambiguity between different systems and achieving global uniformity in meaning, format, and value of multi-source data, laying a high-quality data foundation for subsequent data fusion. By pre-setting the initial fusion granularity based on cross-border business needs and dynamically adjusting the fusion granularity in conjunction with collaboration efficiency parameters, and dynamically allocating fusion weights for multi-dimensional data based on communication quality assessment results and global data accuracy parameters, the optimal balance between information value and processing efficiency in fusion granularity is achieved. This ensures a high degree of adaptability of dynamic data fusion results to business scenarios such as cross-border trade order collaboration, warehouse management, and logistics scheduling, ultimately significantly improving the reliability and efficiency of data collaboration throughout the cross-border trade process, and providing solid data support for intelligent decision-making and refined operation in cross-border trade.

[0027] Furthermore, the communication quality parameters include connection establishment success rate, network disconnection rate, and data accuracy. The steps for performing dynamic communication quality assessment based on the device terminal's communication quality parameters include: obtaining preset communication quality reference parameters and communication quality weighting parameters. The communication quality reference parameters include a connection establishment success rate threshold, a network disconnection rate threshold, and a data accuracy threshold. The communication quality weighting parameters include a connection establishment weighting ...

[0028] The communication quality assessment index is obtained as follows:

[0029] ;

[0030] In the formula, This represents a communication quality assessment index. , and These represent the weighting percentages for connection establishment, network connectivity, and data accuracy, respectively. , and These represent connection establishment success rate, connection network disconnection rate, and data accuracy rate, respectively. , and These represent the connection establishment success rate threshold, the connection network disconnection rate threshold, and the data accuracy rate threshold, respectively. The connection establishment success rate is the proportion of the number of times the terminal successfully accesses the communication network out of the total number of requests. The connection network disconnection rate is the proportion of the number of times the terminal abnormally disconnects from the network out of the total number of requests per unit time. The data accuracy rate is the consistency between the data parsed by the receiving end and the original data collected by the terminal.

[0031] In this embodiment, the present invention constructs a three-dimensional communication quality parameter system encompassing connection establishment success rate, network disconnection rate, and data accuracy. Combined with a two-layer quantitative evaluation mechanism using preset reference thresholds and weighted percentage parameters, it achieves accurate and dynamic evaluation of device terminal communication quality. By selecting three core parameters that directly reflect the entire data transmission process of the device terminal in cross-border trade scenarios, it covers the initial connection stage when the device accesses the network, the link stability stage during data transmission, and the integrity verification stage after data reception, ensuring the comprehensiveness and relevance of the communication quality evaluation. By comparing the actual parameters of the device terminal with preset reference thresholds and then using differentiated weighted percentage parameters to assign weights to each comparison result, it can accurately assess the communication quality of cross-border business operations. Prioritizing different dimensions of communication quality requirements, this approach precisely quantifies the impact of each parameter on overall communication quality, effectively distinguishing the contributions of connection success rate, link stability, and data accuracy in the transmission of different device data types (such as high-priority control commands and high-volume media data). Finally, a communication quality assessment index is obtained through the weighted summation of multi-dimensional influence indices, achieving a quantitative representation of the comprehensive reliability of data transmission between terminal devices and the network. This not only provides precise decision-making basis for subsequent dynamic adjustment of link coding rates and local area link switching, but also effectively avoids link control errors caused by the one-sidedness of single-parameter evaluation, ensuring the stability and reliability of data transmission from cross-border trade equipment terminals and providing high-quality data source support for multi-source data fusion.

[0032] Furthermore, the steps for dynamic link control based on device data type and communication quality dynamic evaluation results include: Device data types include Class I devices, Class II devices, and Class III devices. Class I devices generate small amounts of instruction and control data with high real-time requirements, such as scheduling terminals, controllers, and emergency stop switches. They generate high-priority control instructions and require the lowest network latency and the highest reliability. Class II devices generate large amounts of streaming media data with medium real-time requirements, such as video surveillance cameras and industrial cameras. They generate continuous large volumes of data and require high bandwidth, but allow for a certain amount of latency and jitter. Class III devices generate small amounts of periodic status data with medium real-time requirements, such as temperature sensors, pressure sensors, and power meters. They periodically report small data packets and have relatively relaxed requirements for bandwidth and real-time performance, but require a stable connection. Obtain preset communication quality assessment first threshold and communication quality assessment second threshold; if the communication quality assessment index of any device terminal reaches the communication quality assessment second threshold, it is marked as first communication quality; if the communication quality assessment index of any device terminal reaches the communication quality assessment first threshold but does not reach the communication quality assessment second threshold, it is marked as second communication quality; if the communication quality assessment index of any device terminal does not reach the communication quality assessment first threshold, it is marked as third communication quality; perform dynamic adjustment of link coding rate based on the communication quality of each device terminal and the corresponding device data type; obtain the average communication quality assessment index of each work area based on the communication quality assessment index of each device terminal and each work area; perform local area link handover based on the terminal online rate and average communication quality assessment index of each work area.

[0033] In this embodiment, the present invention achieves intelligent optimization of data transmission links for equipment terminals in cross-border trade scenarios through technical design including equipment data type classification, communication quality threshold classification, dynamic control of link coding rate, and linkage with local area link switching. By classifying equipment terminals into three categories based on data real-time performance and data volume, and clarifying the data transmission requirements of different types of equipment, the invention can accurately match the differentiated data transmission requirements of cross-border trade warehousing, loading and unloading, and transportation operations, avoiding high-priority data transmission delays or bandwidth resource waste caused by indiscriminate link control. By setting two levels of communication quality assessment thresholds and classifying equipment communication quality into three levels, the invention achieves refined and quantitative judgment of the communication status of equipment terminals, providing a clear and actionable decision-making basis for subsequent link control. Through dynamic control of the link coding rate based on equipment communication quality level and data type, the invention can specifically adjust... The encoding rate parameters of different devices ensure low-latency and high-reliability transmission of data from the first type of high-priority control devices, meet the high-bandwidth requirements of the second type of high-volume media devices, and also take into account the stable connection requirements of the third type of periodic state devices. By switching local links based on the average communication quality assessment index of the work area and the terminal online rate, the allocation of link resources can be coordinated at the regional level, avoiding the impact of communication failures of a single device on the overall data transmission of the work area. This effectively improves the stability and efficiency of data transmission in cross-border trade multi-device collaborative scenarios, laying a solid link foundation for the high-quality access and integration of subsequent multi-source data.

[0034] Furthermore, the steps for dynamically adjusting the link coding rate based on the communication quality of each device terminal and the corresponding device data type include: matching the device data type with a preset coding rate mapping table and a coding rate adjustment unit mapping table to obtain the corresponding coding rate standard range and coding rate adjustment unit. The coding rate mapping table specifies the coding rate standard range adopted for different device data types, and the coding rate adjustment unit mapping table specifies the coding rate adjustment unit adopted for different device data types. If the communication quality of the device terminal is the first communication quality, the difference between the communication quality evaluation index and the second communication quality evaluation threshold is marked as the communication quality surplus. The surplus index is used to match the communication quality surplus index with a preset coding rate first adjustment multiple mapping table to obtain the corresponding coding rate adjustment multiple. The coding rate first adjustment multiple mapping table specifies the coding rate adjustment multiple to be used under the first communication quality for different communication quality surplus index ranges. The current link coding rate is dynamically adjusted based on the product of the coding rate adjustment multiple and the coding rate adjustment unit corresponding to the device data type. The current link coding rate is then processed based on the product of the coding rate adjustment multiple and the coding rate adjustment unit corresponding to the device data type to obtain the target coding rate. If the target coding rate exceeds the coding rate of the corresponding device data type... If the maximum value of the standard range for the code rate is reached, the coding rate is directly adjusted to the maximum value of the standard range for the corresponding device data type; otherwise, the coding rate is adjusted to the target coding rate. If the communication quality of the device terminal is the second communication quality, the difference between the second threshold for communication quality assessment and the communication quality assessment index is marked as the communication quality deviation index. The communication quality deviation index is matched with a preset second coding rate adjustment multiple mapping table to obtain the corresponding coding rate adjustment multiple. The second coding rate adjustment multiple mapping table specifies the coding rate adjustment ratio to be adopted under the second communication quality for different communication quality deviation index ranges. Based on the device data type... The product of the coding rate adjustment unit and the coding rate adjustment multiple is summed with the current link coding rate to obtain the adjusted coding rate. If the communication quality of the device terminal is the third communication quality, the corresponding coding rate adjustment multiple is obtained based on the communication quality deviation index. The product of the coding rate adjustment multiple and the coding rate adjustment unit corresponding to the device data type is subtracted based on the current link coding rate to obtain the target coding rate. If the target coding rate is lower than the minimum value of the standard range of coding rate for the corresponding device data type, the coding rate is adjusted to the minimum value of the standard range of coding rate for the corresponding device data type; otherwise, the coding rate is adjusted to the target coding rate.

[0035] In this embodiment, the present invention uses a preset coding rate mapping table and a coding rate adjustment unit mapping table to match differentiated coding rate standard ranges and adjustment granularities for devices of different data types. This satisfies the low-latency and high-reliability transmission requirements of the first type of high-real-time control devices, adapts to the high-bandwidth requirements of the second type of high-volume media devices, and also takes into account the stable connection requirements of the third type of periodic state devices, avoiding resource waste or transmission performance degradation caused by indiscriminate coding rate configuration. By designing differentiated adjustment logic for different communication quality levels, the coding rate is positively increased based on the communication quality surplus index under the first communication quality level to fully utilize link resources. Under the second communication quality condition, the coding rate is finely adjusted based on the communication quality deviation index to maintain transmission stability. Under the third communication quality condition, the coding rate is reduced inversely based on the communication quality deviation index to ensure data transmission integrity, thus achieving dynamic adaptation between the coding rate and the communication quality state. At the same time, by setting upper and lower limits of the coding rate standard range, it is ensured that the adjusted coding rate is always within the optimal range for device data transmission, effectively avoiding data packet loss caused by excessively high coding rates or transmission delays caused by excessively low coding rates. This significantly improves the adaptability and efficiency of data transmission for various types of devices under different communication quality conditions, providing a stable link transmission guarantee for high-quality access to multi-source data in cross-border trade.

[0036] Furthermore, the steps for performing local area link switching based on the terminal online rate and average communication quality assessment index of each work area include: if the average communication quality assessment index of any work area is lower than the first threshold of communication quality assessment, a work area link investigation is triggered. First, a link layered diagnostic mechanism is activated, and the physical connection status between each device terminal and the link access point, the operating parameters of the link transmission protocol, and the routing configuration of the data forwarding node are investigated in the order of terminal access layer, transmission network layer, and data forwarding layer. At the terminal access layer, the link access authentication status, signal strength, and port working mode of each device terminal are checked to locate terminal devices that are not connected or have abnormal access. At the transmission network layer, the bandwidth utilization of the link is collected. Key indicators such as throughput, packet loss rate, and transmission delay are analyzed to determine if there are link congestion or signal interference issues. At the data forwarding layer, the validity of the routing table, the matching degree of the forwarding strategy, and the load status of node devices are checked to identify data forwarding bottlenecks. Simultaneously, the historical operation logs of the links in the work area are retrieved, and the current link status is compared with the historical normal status to pinpoint the time point of the anomaly and its evolution trend. Finally, the investigation results at each level are summarized to generate a link anomaly diagnosis report, clarifying the cause of the anomaly and the corresponding repair plan, providing data support for ensuring stability after link switching. The links in the work area are switched to backup links, and the monitoring frequency of the work area is adjusted to the preset maximum value. If the average throughput of any work area... If the communication quality assessment index is not lower than the first threshold but lower than the second threshold, the link in the work area will be switched to a load-sharing link. Specifically, non-critical data will be diverted to the main transport link, and critical data will be diverted to the secondary transport link. If the average communication quality assessment index of the work area is still lower than the second threshold after the switch, the link will be switched directly to the backup link, and it will be determined whether the terminal online rate has reached the preset online rate threshold. If so, no additional processing will be performed; otherwise, the difference between the online rate threshold and the terminal online rate will be marked as the online deviation rate. Based on the online deviation rate, the monitoring frequency of the work area will be dynamically increased. The preset benchmark monitoring frequency and the online deviation rate will be weighted to obtain the enhanced target. The monitoring frequency is positively correlated with the online deviation rate and the monitoring frequency adjustment coefficient. The higher the online deviation rate, the higher the corresponding monitoring frequency adjustment coefficient. Based on the product of the benchmark monitoring frequency and the monitoring frequency adjustment coefficient, the target monitoring frequency for the work area is determined. The monitoring task scheduling strategy for the work area is updated synchronously. The communication quality parameters, link connection status and data transmission rate of all equipment terminals in the work area are collected periodically according to the target monitoring frequency. The collected real-time data is compared and analyzed with preset thresholds. Once abnormal fluctuations occur in the monitoring data, an early warning mechanism is immediately triggered and an abnormal log is recorded. This achieves dynamic adaptation of the monitoring frequency to the degree of deviation of the terminal online rate, ensuring refined management and control of work areas with low online rates.If the average communication quality assessment index of any work area is not lower than the second threshold of communication quality assessment, the current primary link will be maintained.

[0037] The primary link is the core link that carries all data transmissions by default in the work area. It features high bandwidth, low latency, and high stability, and is the preferred channel for transmitting data from cross-border trade equipment terminals and ERP business data. Under normal conditions, it undertakes the transmission of all data. The backup link is a redundant backup path for the primary link. It is deployed to deal with scenarios where the primary link fails or the communication quality deteriorates. It is usually in standby or low-load state and is only switched on and activated when the primary link cannot meet the transmission requirements, ensuring the continuity of data transmission. The primary transmission link, in load-sharing mode, is a link dedicated to carrying critical data. Its underlying layer can be the currently active primary link. Its core function is to ensure the reliability of high-priority data transmission. Critical data includes: equipment control commands, cross-border order fulfillment data, customs declaration core data, real-time inventory scheduling data, etc. This type of data is extremely sensitive to latency and packet loss rate, directly affecting the normal operation of cross-border business. The secondary transmission link, in load-sharing mode, is a link dedicated to carrying non-critical data. Its underlying layer can be the remaining bandwidth resources of the primary link or a low-priority path running parallel to the primary link. Its core function is to share the load pressure of the primary transmission link. Non-critical data includes: periodic equipment status reporting data, system log data, historical business statistics data, etc. This type of data has lower requirements for transmission timeliness and allows for a certain degree of latency and bandwidth compression.

[0038] In this embodiment, the present invention establishes a dual judgment standard of average communication quality assessment index and terminal online rate, combined with a hierarchical link investigation mechanism, to achieve full-dimensional positioning and root cause tracing of link anomalies. This includes device status verification at the terminal access layer, link performance collection at the transmission network layer, and node load analysis at the data forwarding layer, along with historical log comparison. This completely solves the problems of the one-sidedness and blindness of traditional link investigation, providing accurate decision-making basis for link switching. Simultaneously, it switches abnormal operation areas to backup links and maximizes monitoring frequency to ensure the continuity and security of data transmission under high-risk links. For operation areas with critical communication quality, a load-sharing link switching strategy is adopted, and differentiated diversion of critical and non-critical data is implemented. This ensures the transmission priority of critical data such as cross-border order control instructions and core customs declaration data, while avoiding the main link being overloaded through reasonable diversion of non-critical data. Congestion was addressed, enabling refined allocation of link resources. A proactive prevention and control system for low-online-rate work areas was constructed through a dynamic adjustment mechanism that correlates online deviation rate with monitoring frequency. The adaptive logic of higher monitoring frequency for higher online deviation rates allows for real-time capture of terminal access status fluctuations. Combined with anomaly warnings and log recording, this effectively mitigates the risk of data transmission interruptions due to terminal offline status, achieving refined control over vulnerable work areas. A tiered processing strategy of maintaining high-quality links, optimizing critical links, and switching inferior links avoids excessive consumption of link resources. While ensuring the reliability of data transmission across all stages of cross-border trade warehousing, loading and unloading, and transportation, it improves the utilization efficiency of link resources. Ultimately, this forms a link management capability adapted to the complex network environment of cross-border trade, providing solid link support for stable access and efficient integration of multi-source data, and significantly reducing the risk of cross-border business delays caused by link problems.

[0039] Furthermore, the collaborative efficiency parameters include inventory turnover days deviation rate, order fulfillment rate, and on-time delivery rate. The steps for dynamically adjusting the fusion granularity based on cross-border business needs and collaborative efficiency parameters include: business demand types include real-time management needs, regional management needs, and operational management needs; if the business demand type is real-time management, the initial data fusion granularity is preset to a fine-grained reference value, characterized by data aggregation for individual devices, with time windows of minutes or hours, to support real-time monitoring and scheduling of device status; if the business demand type is regional management, the initial data fusion granularity is preset to a medium-grained reference value, characterized by aggregation for device clusters within a specific operational area, with time windows of minutes or hours, to support real-time monitoring and scheduling of device status; if the business demand type is regional management, the initial data fusion granularity is preset to a medium-grained reference value, characterized by aggregation for device clusters within a specific operational area, with time windows of minutes or hours, to support real-time monitoring and scheduling of device status. Data aggregation is performed on a daily or weekly time window to support regional collaboration and performance evaluation. If the business requirement is for operation and management, the initial data fusion granularity is preset to a coarse-grained reference value. This is characterized by data aggregation performed on a monthly or quarterly time window for a cross-regional global device network to support systematic operation planning and strategic decision-making. The initial data fusion granularity is dynamically adjusted based on the collaboration efficiency parameter to obtain a data fusion granularity value. This value is then matched with a preset fusion granularity mapping table to obtain the corresponding fusion granularity. The fusion granularity includes fine-grained, medium-grained, and coarse-grained granularities. The fusion granularity mapping table specifies the corresponding fusion granularity to be used for different data fusion granularity ranges. Fine-grained data aggregation focuses on individual devices, with time windows measured in minutes or hours. The aggregated data dimensions include detailed information such as the device's real-time operating status, operation records for each task, and real-time data flow for each order. Its core function is to support real-time monitoring of device status, rapid response to abnormal commands, and precise traceability of order fulfillment in cross-border trade operations. It is suitable for real-time management scenarios such as equipment fault diagnosis and real-time inventory scheduling. Medium-grained data aggregation focuses on clusters of devices within a specific operational area, with time windows measured in days or weeks. The aggregated data dimensions include the total operational volume of all devices within the area, weekly trends in regional inventory, and overall order fulfillment within the region. The first level, summarizing information such as order fulfillment rate, primarily supports regional-level collaborative efficiency assessment, optimized allocation of regional resources, and identification of regional business bottlenecks in cross-border trade. It is suitable for regional management needs such as overseas warehouse operational efficiency analysis and regional logistics link optimization. The second level, coarse-grained, is an overview-level data aggregation granularity for the global equipment network across regions. The time window is on a monthly or quarterly basis. The aggregated data dimensions cover macro-statistics such as the overall inventory turnover of the entire cross-border trade network, global order fulfillment rate, and cross-regional delivery timeliness rate. Its core function is to support the systematic operation planning, long-term strategy formulation, and optimized layout of the entire supply chain in cross-border trade. It is suitable for operational management needs such as annual business target setting and cross-border market expansion decisions.

[0040] In this embodiment, the present invention categorizes cross-border business needs into three types: real-time management, regional management, and operational management. It pre-sets differentiated initial fusion granularities for different needs, satisfying both the real-time management requirement for instant monitoring and scheduling of minute / hourly data aggregation for individual devices, and the regional management requirement for evaluating the collaborative performance of daily / weekly data aggregation for equipment clusters in the work area. Simultaneously, it addresses the strategic planning and decision-making requirements of monthly / quarterly data aggregation for the entire equipment network under operational management, thus resolving the pain point that traditional fixed fusion granularities cannot accommodate multiple types of business needs. Furthermore, by introducing three types of collaborative performance metrics—inventory turnover days deviation rate, order fulfillment rate, and on-time delivery rate—the invention achieves this. The rate parameter dynamically corrects the initial fusion granularity, upgrading the fusion granularity from passively adapting to needs to proactively responding to business efficiency. It can adjust the fineness of data aggregation in real time according to the fluctuations of core business indicators in cross-border trade. For example, when the inventory turnover days deviation rate exceeds the standard, the fusion granularity is automatically refined to achieve accurate monitoring of inventory status. When the order fulfillment rate and on-time delivery rate meet the standards, the fusion granularity is maintained or coarsened to improve data processing efficiency. Ultimately, it achieves the optimal balance between information value and processing efficiency, significantly improving the support capability of multi-source data fusion results for the entire cross-border trade business process, and providing high-quality data decision-making basis for core links such as warehouse scheduling, order fulfillment, and supply chain optimization.

[0041] Furthermore, the step of dynamically adjusting the initial data fusion granularity based on the collaborative efficiency parameters to obtain the data fusion granularity value includes: obtaining preset collaborative efficiency reference parameters, including an inventory turnover days deviation rate threshold, an order fulfillment rate threshold, and a delivery on-time rate threshold; determining whether the inventory turnover days deviation rate exceeds the inventory turnover days deviation rate threshold; if so, marking the difference between the inventory turnover days deviation rate threshold and the inventory turnover days deviation rate as the days performance deviation; matching the days performance deviation with a preset fusion granularity first adjustment mapping table to obtain the corresponding granularity first adjustment value; otherwise, adjusting the granularity... The first adjustment value is marked as 0. The fusion granularity first adjustment mapping table specifies the corresponding granularity first adjustment value to be used for different ranges of performance deviation over different days. The larger the deviation, the larger the adjustment value, indicating that finer data granularity is needed for monitoring. It then checks whether the order fulfillment rate has reached the order fulfillment rate threshold. If so, the granularity second adjustment value is marked as 0; otherwise, the difference between the order fulfillment rate threshold and the order fulfillment rate is marked as the order performance deviation. The order performance deviation is matched with the preset fusion granularity second adjustment mapping table to obtain the corresponding granularity second adjustment value. The fusion granularity second adjustment mapping table specifies the granularity second adjustment value for different orders. The corresponding granularity of the second adjustment value should be used for the performance deviation range, and the larger the deviation, the larger the adjustment value; determine whether the on-time delivery rate has reached the on-time delivery rate threshold. If so, mark the granularity of the third adjustment value as 0; otherwise, mark the difference between the on-time delivery rate threshold and the on-time delivery rate as the delivery performance deviation. Match the delivery performance deviation with the preset fused granularity of the third adjustment mapping table to obtain the corresponding granularity of the third adjustment value. The fused granularity of the third adjustment mapping table specifies the corresponding granularity of the third adjustment value to be used for different delivery performance deviation ranges, and the larger the deviation, the larger the adjustment value; the day performance deviation... The sum of order performance deviation and delivery performance deviation is denoted as the overall deviation. The overall deviation is calculated by comparing the day performance deviation, order performance deviation, and delivery performance deviation with the overall deviation to obtain the collaboration efficiency ratio parameter. This parameter includes the weighting ratios of day performance, order performance, and delivery performance. The corresponding first, second, and third granularity adjustment values ​​are weighted using the collaboration efficiency ratio parameter and then summed to obtain the granularity adjustment value. This granularity adjustment value is then used to correct the initial data fusion granularity, resulting in the data fusion granularity value.

[0042] The data fusion granularity value is obtained as follows:

[0043] ;

[0044] In the formula, Indicates the granularity of data fusion. Indicates the initial data fusion granularity. , and These represent the weighting of collaborative efficiency parameters, including the weighting of days performance, order performance, and delivery performance. , and These represent the first, second, and third granularity adjustment values, respectively. The inventory turnover days deviation rate is calculated based on historical inventory and sales data recorded in the ERP business system. Using a set calculation period as a benchmark, the average inventory value and sales cost within that period are first determined. Then, the average inventory value is multiplied by the number of days in the calculation period, and the result is divided by the sales cost of the calculation period to obtain the theoretical inventory turnover days for cross-border trade within that period. For the same time period as the theoretical value calculation, actual inventory turnover records are extracted and organized from the ERP business data to determine the actual inventory turnover days within that period. The absolute value of the difference between the actual and theoretical inventory turnover days is divided by the theoretical inventory turnover days and then multiplied by 100% to obtain the inventory turnover days deviation rate, used to quantify the degree of deviation in inventory turnover efficiency. The order fulfillment rate is obtained using order fulfillment records in the ERP business data as the core data source, following a logical process of data statistics and proportional calculation. Within this period, all valid orders conforming to business specifications are screened from the order fulfillment records of the ERP system, and the total number of valid orders is calculated. Then, based on preset delivery cycle standards and order demand quantity standards, each valid order is checked individually, and orders that simultaneously meet the requirements of completing the entire delivery process within the preset delivery cycle and whose actual delivery quantity is completely consistent with the quantity agreed upon in the order are defined as on-time fulfillment orders, and their quantity is counted. Finally, the number of on-time fulfillment orders is divided by the total number of valid orders, and the result is multiplied by 100% to obtain the order fulfillment rate, which reflects the quality of cross-border trade order fulfillment. The on-time delivery rate is obtained by combining order delivery records and corresponding logistics tracking data from the ERP business data, through order screening and ratio calculation. First, the calculation period is determined. Within this period, all order information that has completed the entire delivery process is extracted from the order delivery records and logistics tracking data of the ERP system, and the total number of completed delivery orders is calculated. Next, based on the preset allowable deviation range for delivery time, the actual delivery time of each completed order is compared with the agreed delivery time in the order. Orders whose difference between the actual delivery time and the agreed delivery time is within the preset allowable range are selected and defined as on-time delivery orders, and their number is counted. Finally, the number of on-time delivery orders is divided by the total number of completed orders, and the result is multiplied by 100% to obtain the on-time delivery rate, which reflects the accuracy of delivery time in cross-border trade.

[0045] In this embodiment, the present invention accurately quantifies the performance deviation of each parameter by comparing three core collaborative efficiency parameters—inventory turnover, order fulfillment, and delivery timeliness—with preset thresholds and matching corresponding granularity adjustment values. This avoids the one-sidedness of granular correction caused by single-parameter evaluation and ensures that the fusion granularity is tilted towards weak links in the business through a mapping logic where the larger the deviation, the larger the adjustment value. When the inventory turnover days deviation rate exceeds the standard, the granularity is fined to strengthen inventory monitoring; when the order fulfillment rate and delivery timeliness rate are insufficient, the data aggregation accuracy is improved to locate fulfillment bottlenecks, solving the pain point that traditional fixed granularity cannot respond to dynamic changes in business. By using the proportion of each performance deviation to the overall deviation as the collaborative efficiency ratio parameter, the invention achieves... The dynamic allocation of adjustment value weights enables granularity correction to focus on the performance indicators that have the greatest impact on cross-border business, avoiding resource waste caused by indiscriminate weighting. Finally, the initial fusion granularity is corrected by the granularity adjustment value obtained by weighted summation. This ensures that the data fusion granularity value not only matches the initial positioning of different business needs such as real-time management and regional management, but also accurately responds to fluctuations in business performance. This achieves dynamic optimization of fusion granularity based on business needs and performance data, ensuring that subsequent data fusion results can provide sufficient information support for scenarios such as equipment scheduling and regional collaboration, while avoiding processing efficiency losses caused by excessive refinement. This provides a more adaptable data foundation for efficient collaboration and intelligent decision-making throughout the entire cross-border trade process.

[0046] Figure 2This document presents a dynamic data fusion flowchart for an ERP multi-source data fusion method for cross-border trade provided in this application embodiment. The critical threshold in the flowchart represents the communication quality critical threshold, and the fusion weight is set to 0. The steps for performing dynamic data fusion based on the overall data accuracy parameter, the dynamic adjustment result of the fusion granularity, and the dynamic evaluation result of the communication quality include: if the communication quality of the device terminal is the first communication quality, then the fusion weight of the device terminal data is set to a preset first weight, indicating that the device's communication link is stable and reliable, and its data should dominate the fusion process; if the communication quality of the device terminal is the second communication quality, then the fusion weight of the device terminal data is set to a preset second weight, indicating that the device's communication has certain fluctuations, and its data can participate in the fusion, but its reliability needs to be appropriately discounted; if the communication quality of the device terminal is the third communication quality, then it is determined whether the communication quality evaluation index reaches the preset communication quality critical threshold. If so, the fusion weight of the device terminal data is set to a preset third weight, indicating that the device's communication quality is poor, and the data may be unreliable. In this case, its weight can be set to extremely low to reduce its impact on the fusion process. The impact is negligible; otherwise, the fusion weight of the device terminal data is set to 0, indicating that the data of this device is excluded from this fusion due to poor quality, and triggering a device anomaly warning. The fusion weight of ERP business data and the fusion weight of external related data are dynamically obtained based on the fusion weight of device terminal data and the accuracy parameter of the whole domain data. Multi-dimensional data is dynamically aggregated based on the data fusion granularity. If the fusion granularity is fine-grained, multi-dimensional data is aggregated for a single device, with a window of minutes or hours, to form device-level detailed data. If the fusion granularity is medium-grained, multi-dimensional data is aggregated for a device cluster in a specific work area, with a window of days or weeks, to form regional summary data. If the fusion granularity is coarse-grained, multi-dimensional data is aggregated for the global device network, with a window of months or quarters, to form system-level overview data. The aggregated multi-dimensional data is then weighted and fused with the corresponding fusion weights to obtain the final fusion result.

[0047] In this embodiment, the present invention effectively solves the core problems of difficult data reliability control, rigid weight allocation, and disconnect between fusion results and business operations in cross-border trade ERP multi-source data fusion. Based on communication quality assessment results, it implements hierarchical weight allocation for device terminal data. A high weight under the first communication quality ensures the dominant position of stable link data; a discounted weight under the second communication quality balances the participation value and risk of fluctuating data; and a very low or zero weight under the third communication quality filters unreliable data and triggers early warnings. This achieves differentiated management of device terminal data, with higher weights for superior data and constraints for inferior data, thus improving the basic reliability of fused data from the source. Furthermore, it dynamically allocates the fusion weights of ERP business data and externally related data according to the accuracy parameters of the overall data, closely linking the weight ratio to the ontological credibility of the two types of data, avoiding fixed weights. This addresses the issues of insufficient weighting of credible data and interference from questionable data, ensuring the accuracy of multi-source data fusion. Finally, by dynamically adjusting the fusion granularity, multi-dimensional data is aggregated in an adaptive manner. Fine-grained device-level detailed aggregation supports real-time scheduling, medium-grained regional-level summary aggregation serves collaborative evaluation, and coarse-grained system-level overview aggregation assists strategic decision-making. The final result is generated through weighted fusion operations, achieving end-to-end optimization of data aggregation accuracy to meet business needs, fusion weights to match data quality, and result output to adapt to application scenarios. This not only ensures the stringent data reliability requirements of core links such as cross-border trade order fulfillment and inventory scheduling, but also improves data fusion efficiency through differentiated weights and granularity adaptation, providing high-quality and highly adaptable data support for intelligent collaboration and accurate decision-making throughout the entire cross-border trade process.

[0048] Furthermore, the overall data accuracy parameter consists of ERP data accuracy and external related data accuracy. The steps for dynamically obtaining the ERP business data fusion weight and external related data fusion weight based on the fusion weight of device terminal data and the overall data accuracy parameter include: marking the fusion weight of device terminal data with the complement of 1 as the remaining data fusion weight; assigning weights to ERP data accuracy and external related data accuracy using preset ERP data weight ratios and related data weight ratios respectively to obtain overall data accuracy; marking the ratio of the product of ERP data accuracy and ERP data weight ratio to overall data accuracy as ERP business data ratio; marking the ratio of the product of external related data accuracy and related data weight ratio to overall data accuracy as external related data ratio; and marking the product of ERP business data ratio and external related data ratio with the remaining data fusion weight as ERP business data fusion weight and external related data fusion weight respectively.

[0049] In this embodiment, the present invention defines the weight allocation boundary of ERP business data and external related data in the overall fusion by defining the device terminal data fusion weight plus the complement of 1 as the remaining data fusion weight. This ensures the integrity and rationality of the total weight of multi-source data fusion and avoids the problem of one type of data being overly dominant or marginalized due to unbalanced weight allocation. The invention introduces preset ERP data weight ratios and related data weight ratios to perform weighted calculations on the accuracy of the two types of data to generate overall data accuracy. This takes into account the inherent value differences between core ERP business data and external related data in cross-border trade scenarios, and transforms the core quality indicator of data accuracy into a quantifiable overall evaluation standard through weighting, providing an objective basis for subsequent weight allocation. The ratio of the product of data accuracy and its corresponding weight ratio to the overall data accuracy is used as the weight allocation of the two types of data. The final fusion weight is obtained by multiplying the baseline weight by the remaining data fusion weights. This ensures that the weight allocation of ERP business data and external related data is directly linked to their own accuracy and dynamically complements the weights of equipment terminal data. When the accuracy of ERP data is higher, its proportion in the remaining weights naturally increases, and the same applies to external related data. This completely solves the pain point that traditional fixed weight allocation cannot adapt to dynamic changes in data quality, ensuring that high-quality data occupies a more reasonable weight share in the fusion process. At the same time, the accurate calculation of weight proportions avoids the problem of data quality and weight mismatch. Ultimately, it achieves a dynamic balance and optimal configuration of the fusion weights of ERP business data, external related data, and equipment terminal data, making the fusion result more reflective of the true quality level of multi-source data. This provides key weight allocation support for the reliability and business adaptability of multi-source data fusion in cross-border trade ERP.

[0050] like Figure 3The diagram shown is a structural schematic of an ERP multi-source data fusion system for cross-border trade provided in this application embodiment. It includes: a data acquisition module, a link control module, a mapping alignment module, and a data fusion module. The data acquisition module acquires multi-dimensional data to be fused, including ERP business data, external related data, and device terminal data. The link control module performs dynamic communication quality assessment based on the communication quality parameters of the device terminal during cross-border trade data access, and performs dynamic link control based on the device data type and the dynamic communication quality assessment results. Dynamic link control refers to dynamically adjusting the link encoding rate of the device terminal and performing local area link switching in the work area to optimize the reliability and efficiency of data transmission. The mapping alignment module... The alignment module is used to perform heterogeneous difference field evaluation based on the degree of difference in multi-dimensional data, obtain the difference fields of multi-dimensional data, and perform mapping and alignment based on the semantic middleware of multi-dimensional data and the difference fields to eliminate data semantic ambiguity between different systems. The data fusion module is used to perform dynamic adjustment of fusion granularity according to cross-border business needs and collaborative efficiency parameters of multi-dimensional data during the fusion modeling process, and perform dynamic data fusion based on the accuracy parameters of the whole domain data, the results of dynamic adjustment of fusion granularity, and the results of dynamic evaluation of communication quality. Dynamic adjustment of fusion granularity means dynamically adjusting the granularity of data fusion to achieve the optimal balance between information value and processing efficiency. Dynamic data fusion means dynamically adjusting the fusion weight of multi-dimensional data to ensure the reliability and business adaptability of the fusion results.

[0051] The various features and processes described above can be used independently of each other or can be combined in various ways. All possible combinations and sub-combinations are intended to fall within the scope of this disclosure. Furthermore, certain method or process blocks may be omitted in some embodiments. The methods and processes described herein are not limited to any particular order, and the blocks or states associated with them may be performed in other suitable orders. For example, the described blocks or states may be performed in an order different from the order specifically disclosed, or multiple blocks or states may be combined in a single block or state. Example blocks or states may be performed serially, in parallel, or in some other manner. Blocks or states may be added to or removed from the disclosed example embodiments. The exemplary systems and components described herein may be configured differently from those described. For example, elements may be added to, removed from, or rearranged compared to the disclosed example embodiments.

[0052] The embodiments described herein have been described in sufficient detail to enable those skilled in the art to practice the disclosed teachings. Other embodiments may be used and derived therefrom, such that structural and logical substitutions and changes may be made without departing from the scope of this disclosure. Therefore, the detailed description should not be construed as limiting, and the scope of the various embodiments is defined only by the appended claims and the full scope of their equivalents.

Claims

1. An ERP multi-source data fusion method for cross-border trade, characterized in that, Includes the following steps: Acquire multi-dimensional data to be integrated, including ERP business data, external related data, and device terminal data; During the cross-border trade data access process, dynamic communication quality assessment is performed based on the communication quality parameters of the device terminal, and dynamic link control is performed based on the device data type and the dynamic communication quality assessment results. The dynamic link control refers to dynamic control of the link coding rate of the device terminal and local area link switching of the work area to optimize the reliability and efficiency of data transmission. Heterogeneous difference field evaluation is performed based on the degree of difference of the multi-dimensional data to obtain the difference field of the multi-dimensional data. The semantic middleware of the multi-dimensional data and the difference field are mapped and aligned to eliminate data semantic ambiguity between different systems. During the fusion modeling process, dynamic adjustment of fusion granularity is performed based on cross-border business needs and collaborative efficiency parameters of multi-dimensional data. Dynamic data fusion is also performed based on global data accuracy parameters, dynamic adjustment results of fusion granularity, and dynamic evaluation results of communication quality. The dynamic adjustment of fusion granularity means dynamically adjusting the granularity of data fusion to achieve the optimal balance between information value and processing efficiency. The dynamic data fusion means dynamically adjusting the fusion weights of multi-dimensional data to ensure the reliability and business adaptability of the fusion results. The collaborative efficiency parameters include inventory turnover days deviation rate, order fulfillment rate, and on-time delivery rate. The steps for dynamically adjusting the fusion granularity based on the collaborative efficiency parameters of cross-border business needs and multi-dimensional data include: The types of business requirements include real-time management requirements, regional management requirements, and operation management requirements; If the business requirement is a real-time management requirement, the initial data fusion granularity will be preset to a fine-grained reference value. If the business requirement is a regional management requirement, the initial data fusion granularity will be preset to a medium-granularity reference value. If the business requirement type is operation and management requirement, the initial data fusion granularity will be preset to a coarse-grained reference value; The initial data fusion granularity is dynamically corrected based on the collaborative efficiency parameter to obtain the data fusion granularity value. The data fusion granularity value is then matched with a preset fusion granularity mapping table to obtain the corresponding fusion granularity. The fusion granularity includes fine granularity, medium granularity, and coarse granularity.

2. The ERP multi-source data fusion method for cross-border trade according to claim 1, characterized in that: The communication quality parameters include connection establishment success rate, connection network disconnection rate, and data accuracy. The steps for performing dynamic communication quality assessment based on the communication quality parameters of the device terminal include: Obtain preset communication quality reference parameters and communication quality ratio parameters. The communication quality reference parameters include connection establishment success rate threshold, connection network disconnection rate threshold, and data accuracy threshold. The communication quality ratio parameters include connection establishment weight ratio, connection network weight ratio, and data accuracy weight ratio. The connection establishment success rate, connection network disconnection rate threshold, and data accuracy of the device terminal are compared with the connection establishment success rate threshold, connection network disconnection rate threshold, and data accuracy rate threshold, respectively. Then, the comparison results are weighted using the communication quality ratio parameter to obtain the connection establishment success rate impact index, connection network disconnection rate impact index, and data accuracy impact index of the device terminal, respectively. The communication quality assessment index of the device terminal is obtained based on the connection establishment success rate impact index, the connection network disconnection rate impact index, and the data accuracy impact index. The communication quality assessment index represents the overall reliability of data transmission between the terminal device and the network.

3. The ERP multi-source data fusion method for cross-border trade of claim 2, wherein: The steps for performing dynamic link control based on device data type and communication quality dynamic assessment results include: Obtain the preset first threshold and second threshold for communication quality assessment; If the communication quality assessment index of any device terminal reaches the second threshold of communication quality assessment, it is marked as the first communication quality. If the communication quality assessment index of any device terminal reaches the first threshold of communication quality assessment, but does not reach the second threshold of communication quality assessment, it is marked as the second communication quality. If the communication quality assessment index of any device terminal does not reach the first threshold of communication quality assessment, it is marked as third communication quality. Dynamic adjustment of link coding rate is performed based on the communication quality of each device terminal and the corresponding device data type; Based on the communication quality assessment index of each device terminal and each work area, the average communication quality assessment index of each work area is obtained. Local area link handover is performed based on the terminal online rate and average communication quality assessment index of each operating area.

4. The ERP multi-source data fusion method for cross-border trade of claim 3, wherein: The steps for dynamically adjusting the link coding rate based on the communication quality of each device terminal and the corresponding device data type include: The device data types are matched with the preset encoding rate mapping table and encoding rate adjustment unit mapping table respectively to obtain the corresponding encoding rate standard range and encoding rate adjustment unit; If the communication quality of the device terminal is the first communication quality, the difference between the communication quality evaluation index and the second communication quality evaluation threshold is marked as the communication quality surplus index. The communication quality surplus index is matched with the preset coding rate first adjustment multiple mapping table to obtain the corresponding coding rate adjustment multiple. If the communication quality of the device terminal is the second communication quality, the difference between the second threshold for communication quality assessment and the communication quality assessment index is marked as the communication quality deviation index. The communication quality deviation index is matched with the preset second coding rate adjustment multiple mapping table to obtain the corresponding coding rate adjustment multiple. The current link coding rate is dynamically adjusted based on the coding rate adjustment unit and coding rate adjustment multiplier corresponding to the device data type. If the communication quality of the device terminal is the third level, the corresponding coding rate adjustment multiple is obtained based on the communication quality deviation index. The current link coding rate is then dynamically adjusted based on the coding rate adjustment multiple and the coding rate adjustment unit corresponding to the device data type.

5. The ERP multi-source data fusion method for cross-border trade of claim 3, wherein: The steps for performing local area link handover based on the terminal online rate and average communication quality assessment index of each work area include: If the average communication quality assessment index of any work area is lower than the first threshold for communication quality assessment, a link check of the work area will be triggered, the link of the work area will be switched to a backup link, and the monitoring frequency of the work area will be adjusted to the preset maximum value. If the average communication quality assessment index of any work area is not lower than the first threshold of communication quality assessment, but lower than the second threshold of communication quality assessment, then the link of that work area will be switched to the load-sharing link. Specifically, non-critical data will be diverted to the main transport link, and critical data will be diverted to the secondary transport link. If the average communication quality assessment index of the work area is still lower than the second threshold of communication quality assessment after the switch, the switch will be made directly to the backup link, and it will be determined whether the terminal online rate has reached the preset online rate threshold. If so, no additional processing will be performed. Otherwise, the difference between the online rate threshold and the terminal online rate is marked as the online deviation rate, and the monitoring frequency of the work area is dynamically increased based on the online deviation rate; If the average communication quality assessment index of any work area is not lower than the second threshold of communication quality assessment, the current primary link will be maintained.

6. The ERP multi-source data fusion method for cross-border trade of claim 1, wherein: The step of dynamically correcting the initial data fusion granularity based on the collaborative efficiency parameter to obtain the data fusion granularity value includes: Obtain preset collaborative efficiency reference parameters, including inventory turnover days deviation rate threshold, order fulfillment rate threshold, and on-time delivery rate threshold; Determine whether the inventory turnover days deviation rate exceeds the inventory turnover days deviation rate threshold. If so, mark the difference between the inventory turnover days deviation rate threshold and the inventory turnover days deviation rate as the days performance deviation degree. Match the days performance deviation degree with the preset fusion granularity first adjustment mapping table to obtain the corresponding granularity first adjustment value. Otherwise, mark the granularity first adjustment value as 0. Determine whether the order fulfillment rate has reached the order fulfillment rate threshold. If so, mark the second granularity adjustment value as 0. Otherwise, mark the difference between the order fulfillment rate threshold and the order fulfillment rate as the order performance deviation. Match the order performance deviation with the preset fusion granularity second adjustment mapping table to obtain the corresponding granularity second adjustment value. Determine whether the on-time delivery rate has reached the on-time delivery rate threshold. If so, mark the third granularity adjustment value as 0. Otherwise, mark the difference between the on-time delivery rate threshold and the on-time delivery rate as the delivery performance deviation. Match the delivery performance deviation with the preset fusion granularity third adjustment mapping table to obtain the corresponding granularity third adjustment value. Based on the deviation of day performance, order performance deviation, and delivery performance deviation, a collaborative efficiency ratio parameter is obtained, which includes the weight ratio of day performance, the weight ratio of order performance, and the weight ratio of delivery performance. The initial data fusion granularity is dynamically corrected based on the collaborative efficiency ratio parameter, the first granularity adjustment value, the second granularity adjustment value, and the third granularity adjustment value to obtain the data fusion granularity value.

7. The ERP multi-source data fusion method for cross-border trade of claim 1, wherein: The steps for performing dynamic data fusion based on global data accuracy parameters, dynamic adjustment results of fusion granularity, and dynamic evaluation results of communication quality include: If the communication quality of the device terminal is the first communication quality, then the fusion weight of the device terminal data is set to the preset first weight; If the communication quality of the device terminal is the second communication quality, then the fusion weight of the device terminal data is set to the preset second weight; If the communication quality of the device terminal is the third communication quality, then determine whether the communication quality evaluation index has reached the preset communication quality critical threshold. If so, set the fusion weight of the device terminal data to the preset third weight; otherwise, set the fusion weight of the device terminal data to 0 and trigger a device anomaly warning. Dynamically obtain the fusion weight of ERP business data and the fusion weight of external related data based on the fusion weight of device terminal data and the accuracy parameter of full-domain data; Based on the dynamic aggregation of multi-dimensional data at the data fusion granularity, the aggregated multi-dimensional data is weighted and fused with the corresponding fusion weights to obtain the final fusion result.

8. The ERP multi-source data fusion method for cross-border trade of claim 7, wherein: The overall data accuracy parameter refers to the accuracy of ERP data and the accuracy of externally related data. The steps for dynamically obtaining the ERP business data fusion weight and external related data fusion weight based on the fusion weight of device terminal data and the accuracy parameter of full-domain data include: The fusion weight of the device terminal data plus the complement of 1 is marked as the fusion weight of the remaining data; The accuracy of ERP data and the accuracy of external related data are weighted by using preset ERP data weight ratios and related data weight ratios respectively, so as to obtain the overall data accuracy. The product of ERP data accuracy and ERP data weight ratio, and the ratio of this product to the overall data accuracy, are denoted as the ERP business data weight ratio. The ratio of the product of the accuracy of external related data and the weight of related data to the accuracy of the whole data is denoted as the external related data ratio. The product of the proportion of ERP business data and the proportion of external related data, and the fusion weight of the remaining data, are respectively labeled as the ERP business data fusion weight and the external related data fusion weight.

9. The ERP multi-source data fusion system for cross-border trade, applying the ERP multi-source data fusion method for cross-border trade according to any one of claims 1-8, characterized in that, include: Data acquisition module, link control module, mapping alignment module, and data fusion module; The data acquisition module is used to acquire multi-dimensional data to be integrated, including ERP business data, external related data, and device terminal data. The link control module is used to perform dynamic communication quality assessment based on the communication quality parameters of the device terminal during the cross-border trade data access process, and to perform dynamic link control based on the device data type and the dynamic communication quality assessment results. The dynamic link control means performing dynamic control of the link encoding rate of the device terminal and performing local area link switching in the work area to optimize the reliability and efficiency of data transmission. The mapping and alignment module is used to perform heterogeneous difference field evaluation based on the degree of difference of the multi-dimensional data to obtain the difference field of the multi-dimensional data, and perform mapping and alignment based on the semantic middleware of the multi-dimensional data and the difference field to eliminate data semantic ambiguity between different systems. The data fusion module is used to dynamically adjust the fusion granularity based on cross-border business needs and the collaborative efficiency parameters of multi-dimensional data during the fusion modeling process. It also performs dynamic data fusion based on the accuracy parameters of the whole domain data, the results of dynamic adjustment of fusion granularity, and the results of dynamic evaluation of communication quality. The dynamic adjustment of fusion granularity means dynamically adjusting the granularity of data fusion to achieve an optimal balance between information value and processing efficiency. The dynamic data fusion means dynamically adjusting the fusion weights of multi-dimensional data to ensure the reliability and business adaptability of the fusion results.

Images