A method for launching an ERP system

Abstract

The present invention relates to the technical field of information system integration, and in particular to an ERP system launch method. The method comprises the following steps: obtaining ERP demand data; performing ERP related function module analysis according to the ERP demand data to generate ERP related function modules; performing ERP system architecture analysis based on the ERP related function modules to establish the ERP system architecture; obtaining historical enterprise source data; performing historical enterprise decision feature analysis based on the historical enterprise source data to generate historical enterprise decision feature data; performing intelligent business practice engine design based on the historical enterprise decision feature data to generate an intelligent business practice engine; performing intelligent business practice function configuration processing of the system architecture on the ERP system architecture based on the intelligent business practice engine to generate an ERP system, and transmitting the ERP system to a cloud server to execute the ERP system launch operation. The present invention realizes intelligent business management of the ERP system.

Description

Technical Field

[0001] The present invention relates to the technical field of information system integration, and in particular to an online method for an ERP system. Background Art

[0002] As an integrated management information system, the Enterprise Resource Planning (ERP) system aims to effectively integrate the internal and external resources of the enterprise through information technology to achieve the purpose of optimizing resource allocation and efficient management. With the expansion of the business scale of enterprises and the continuous increase in management needs, the application of ERP systems has become the key to improving the management level and competitiveness of enterprises. The implementation and launch of ERP systems involve a lot of technical and management work, including system selection, demand analysis, system configuration, data migration, user training, business process optimization, etc. Due to the complexity of ERP systems, long implementation cycles and high costs, how to successfully launch ERP systems efficiently and with low risks has become a common concern of enterprises. However, in the traditional ERP system launch method, the functional module design of the ERP system is disordered and cannot provide users with intelligent business practice services. Summary of the invention

[0003] Based on this, the present invention provides an ERP system online method to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for launching an ERP system includes the following steps:

[0005] Step S1: Acquire ERP demand data; perform ERP related demand analysis based on the ERP demand data to generate ERP related demand data; perform related demand data aggregation processing on the ERP related demand data to generate ERP related aggregated demand data;

[0006] Step S2: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate ERP-related function modules; ERP system architecture analysis is performed based on the ERP-related function modules to establish the ERP system architecture;

[0007] Step S3: Acquire historical enterprise source data; remove redundant features from the historical enterprise source data to generate historical enterprise feature data; perform historical enterprise decision feature analysis based on the historical enterprise feature data to generate historical enterprise decision feature data;

[0008] Step S4: Analyze the enterprise intelligent decision logic parameters according to the historical enterprise decision feature data to generate the enterprise intelligent decision logic parameters; design the intelligent business practice engine based on the enterprise intelligent decision logic parameters to generate the intelligent business practice engine;

[0009] Step S5: Based on the intelligent business practice engine, the intelligent business practice function configuration processing of the ERP system architecture is performed to generate an ERP system, and the ERP system is transmitted to the cloud server to execute the ERP system online operation.

[0010] The present invention can fully understand the specific needs of enterprises in various business modules by acquiring ERP demand data, provide basic data for subsequent demand analysis, and perform related demand analysis on ERP demand data, so as to accurately identify the related needs between various business modules and avoid omission and duplication of needs. Through the aggregation processing of related demand data, the scattered needs can be integrated, and comprehensive ERP related aggregation demand data can be generated, so as to improve the efficiency and accuracy of demand management. Functional module analysis is performed on ERP related aggregation demand data to generate functional modules for enterprise needs, improve the modularization degree of system design, and facilitate system maintenance and expansion. Through ERP system architecture analysis, the architecture of the system is reasonably planned to ensure that each module of the system can work efficiently and collaboratively, and improve the overall performance and stability of the system. Through redundant feature elimination of historical enterprise source data, useless or repeated data features can be removed, streamlined historical enterprise feature data can be generated, and data quality can be improved. Based on historical enterprise feature data, decision feature analysis is performed, data features valuable to enterprise decision-making can be excavated, historical enterprise decision feature data can be generated, and support can be provided for enterprise intelligent decision-making. Through enterprise intelligent decision logic parameter analysis, accurate decision logic parameters are generated, and scientific basis is provided for enterprise intelligent decision-making, and the accuracy and effectiveness of decision-making are improved. Designing an intelligent business practice engine based on intelligent decision-making logic parameters can integrate decision logic into business practice, form an intelligent business processing engine, and improve the intelligence level of enterprise business processing. Functional configuration processing of the system architecture by the intelligent business practice engine can realize intelligent configuration of system functions, ensuring that the ERP system can flexibly respond to the actual business needs of the enterprise. The generated ERP system is transmitted to the cloud server to perform the online operation, which can utilize the efficiency and flexibility of cloud computing to achieve rapid deployment and online of the ERP system, reducing the implementation cycle and cost. Therefore, an online method of an ERP system of the present invention ensures efficient coordination and integration between system modules through the association and aggregation analysis of functional modules, ensures the effective execution of the functions of the ERP system, and satisfies the user's intelligent experience service through the intelligent business practice engine design.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: Obtain ERP demand data;

[0013] Step S12: using natural language technology to perform ERP demand information mining on standard ERP demand data to generate ERP demand mining data;

[0014] Step S13: performing ERP-related demand analysis on the ERP demand mining data to generate ERP-related demand data;

[0015] Step S14: performing associated demand data aggregation processing on the ERP associated demand data to generate ERP associated aggregated demand data.

[0016] The present invention can comprehensively collect the specific needs of each business module of an enterprise by acquiring ERP demand data, provide basic data for subsequent analysis and processing, and avoid misjudgment of needs due to incomplete or inaccurate data. By using natural language processing technology, valuable information can be efficiently extracted from standard ERP demand data to generate structured ERP demand mining data. Natural language processing technology can understand and parse complex text information, improve the accuracy of demand mining, and avoid errors and omissions in manual analysis. Through the association analysis of ERP demand mining data, the associated needs between different business modules can be identified, the isolated design of each module demand can be avoided, and each demand can be prioritized according to the associated demand data to ensure that key needs are given priority, thereby improving the scientificity and rationality of system design. Through aggregation processing, the scattered associated demand data are integrated to generate comprehensive and comprehensive ERP associated aggregated demand data, which is convenient for the design of subsequent functional modules and the construction of system architecture. Aggregation processing can effectively reduce redundancy and conflict in demand data, ensure the consistency and integrity of demand data, and improve the efficiency and quality of system design.

[0017] Preferably, step S13 comprises the following steps:

[0018] Step S131: Analyze and process ERP demand entity information according to ERP demand mining data to generate ERP demand entity data;

[0019] Step S132: performing ERP demand dependency analysis according to the ERP demand entity data to generate ERP demand dependency data;

[0020] Step S133: using the word frequency-inverse document frequency technology to convert the ERP demand mining data into numerical vectors to generate ERP demand mining vector data;

[0021] Step S134: performing ERP demand similarity analysis based on the ERP demand mining vector data to generate ERP demand similarity data;

[0022] Step S135: performing ERP-related demand analysis on the ERP demand mining data based on the ERP demand dependency data and the ERP demand similarity data to generate ERP-related demand data.

[0023] The present invention can clearly identify each specific demand object through ERP demand entity information analysis, generate ERP demand entity data, improve the accuracy of demand analysis, convert complex demand text data into structured entity data, facilitate subsequent processing and analysis, and improve the efficiency and accuracy of data processing. By analyzing the dependency relationship between ERP demand entity data, the mutual relationship between each demand can be identified, and ERP demand dependency relationship data can be generated, which is helpful for the overall design of the system, clarifies the dependency relationship between demands, facilitates the priority sorting of demands and the reasonable allocation of resources, and improves the scientificity and rationality of demand processing. Through the term frequency-inverse document frequency (TF-IDF) technology, the text data is converted into numerical vector data, the representation form of the demand data is standardized, and the uniformity of data processing is improved. The TF-IDF technology can effectively extract key information in the demand text and improve the accuracy and effectiveness of data analysis. Through similarity analysis, similar parts in the demand data can be identified, and ERP demand similarity data can be generated. Similarity analysis helps to perform association analysis on similar demands, facilitates the integration and optimization of demands, and improves the scientificity of demand management. Comprehensively consider the dependency and similarity data of requirements, conduct a comprehensive correlation analysis on the demand mining data, generate ERP related demand data, ensure the comprehensiveness and accuracy of the demand analysis, and through the correlation demand analysis, provide more accurate demand data support for system design and improve the scientificity and rationality of the system functional module design.

[0024] Preferably, step S2 comprises the following steps:

[0025] Step S21: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate an ERP-related function module;

[0026] Step S22: performing microservice architecture mapping processing of the functional modules according to the ERP-related functional modules to generate microservice architecture mapping data;

[0027] Step S23: Containerize the microservice architecture mapping data to generate containerized microservice architecture mapping data, and establish an ERP system architecture through the containerized microservice architecture mapping data.

[0028] The present invention analyzes the functional modules of ERP associated aggregated demand data, and can reasonably divide the demand data into multiple functional modules, improve the modularity of system design, and the division of functional modules helps to clarify the division of labor, improve the efficiency of system design and development, and facilitate the independent development and testing of modules. Mapping the functional modules to the microservice architecture can improve the flexibility and scalability of the system, so that the system can quickly respond to changes in business needs. The microservice architecture can split the system functions into multiple independent services, improve the reliability and stability of the system, and reduce the impact of single point failures. Through containerization, the consistency of each microservice in different environments can be ensured, the problems caused by environmental differences can be reduced, and the stability of system deployment can be improved. The containerization technology makes the deployment and operation and maintenance of the system simpler and more efficient, and can achieve rapid deployment, expansion and migration, improve the efficiency of system launch and maintenance, provide good resource isolation and management capabilities, and can better utilize system resources and improve the performance and security of the system.

[0029] Preferably, step S3 comprises the following steps:

[0030] Step S31: Obtain historical enterprise source data;

[0031] Step S32: performing cluster analysis on the historical enterprise source data based on a preset clustering algorithm to generate historical enterprise cluster data;

[0032] Step S33: Calculate the clustering silhouette coefficient based on the historical enterprise clustering data to generate the clustering silhouette coefficient;

[0033] Step S34: performing redundant information analysis on the historical enterprise source data according to the clustering silhouette coefficient to generate historical enterprise redundant data;

[0034] Step S35: Eliminate redundant features of historical enterprise source data according to historical enterprise redundant data to generate historical enterprise feature data;

[0035] Step S36: Performing an enterprise decision gain impact benefit analysis based on historical enterprise characteristic data to generate enterprise decision gain impact benefit data;

[0036] Step S37: extracting enterprise decision features of the enterprise decision gain affecting benefits from the historical enterprise feature data according to the enterprise decision gain affecting benefits data, and generating historical enterprise decision feature data.

[0037] The present invention can provide a solid foundation for subsequent data analysis and processing by acquiring historical enterprise source data, ensure the comprehensiveness and accuracy of data, collect and integrate enterprise historical data, help summarize and analyze the past operation and decision-making of enterprises, and provide reference for future decision-making. Through cluster analysis, historical enterprise source data are reasonably grouped to generate historical enterprise cluster data, which is helpful to identify patterns and trends in data. Cluster analysis can effectively reduce the complexity of data and improve the efficiency of data processing and analysis. By calculating the clustering silhouette coefficient, the effect of cluster analysis is evaluated to ensure the rationality and accuracy of clustering results. The clustering silhouette coefficient helps to identify and optimize clustering algorithm parameters and improve the quality of cluster analysis. Through redundant information analysis, redundant information in historical enterprise source data can be identified and eliminated, historical enterprise redundant data is generated, data structure is optimized, data redundancy is reduced, data quality and consistency are improved, and higher quality data support is provided for subsequent data processing and analysis. The elimination of redundant features generates streamlined historical enterprise feature data, reduces the complexity of data processing and storage costs, and after removing redundant features, the data analysis results are more accurate and the decision-making basis is more reliable. The analysis of the gain-effect benefit of historical enterprise characteristic data can evaluate the benefits of enterprise decisions, generate enterprise decision-making gain-effect benefit data, provide scientific decision-making basis for enterprises, help enterprises identify high-efficiency decision-making models, optimize the decision-making process, and improve the overall benefits of enterprises. The analysis of the gain-effect benefit data of enterprise decisions can extract the key features that have an important impact on enterprise decisions, generate historical enterprise decision-making feature data, improve the scientificity and effectiveness of enterprises in future decisions, and enhance the competitiveness of enterprises.

[0038] Preferably, step S32 includes the following steps:

[0039] Step S321: Calculate the intra-cluster distance based on the historical enterprise clustering data to generate the intra-cluster distance data;

[0040] Step S322: Calculate the inter-cluster distance based on the historical enterprise clustering data to generate inter-cluster distance data;

[0041] Step S323: Calculate the clustering silhouette coefficient based on the intra-cluster distance data and the inter-cluster distance data to generate the clustering silhouette coefficient.

[0042] The present invention calculates the distance within the cluster, evaluates the closeness of the data points within the same cluster, generates the distance data within the cluster, helps to understand the internal consistency of each cluster, and the calculation result of the distance within the cluster helps to optimize the clustering algorithm, ensures that the data points within the same cluster are closer, thereby improving the accuracy of cluster analysis. By calculating the distance between clusters, the independence between different clusters is evaluated, and the distance data between clusters is generated to ensure the separation and non-overlap of the clustering results. The calculation result of the distance between clusters can be used to improve the clustering algorithm, so that the distance between different clusters is larger, and the effect and accuracy of cluster analysis are improved. Through the calculation of the clustering silhouette coefficient, the quality of the clustering results can be comprehensively evaluated, the clustering silhouette coefficient is generated, and the quality of the clustering effect is understood as a whole. The clustering silhouette coefficient provides a basis for the adjustment of the clustering algorithm, helps to identify problems in the clustering results, and performs corresponding optimization and improvement to improve the quality and reliability of cluster analysis.

[0043] Preferably, step S36 includes the following steps:

[0044] Step S361: Evaluate the impact of historical enterprise decisions on benefits based on historical enterprise characteristic data to generate historical enterprise decision impact benefit data;

[0045] Step S362: Perform enterprise decision gain impact benefit analysis based on historical enterprise decision impact benefit data to generate enterprise decision gain impact benefit data.

[0046] The present invention evaluates historical enterprise characteristic data, can understand the actual impact and benefits of past enterprise decisions, generate historical enterprise decision-making impact benefit data, and help enterprises summarize past decision-making experiences and lessons. The historical enterprise decision-making impact benefit data provides an important reference basis for future decisions, helping enterprises better understand which decisions have a positive impact on enterprise benefits and which decisions may have problems. The analysis of historical enterprise decision-making impact benefit data can identify the decisions that have the greatest impact on enterprise gains, generate enterprise decision-making gain impact benefit data, and help enterprises optimize the decision-making process and improve the scientificity and accuracy of decisions. Enterprise decision-making gain impact benefit data can help enterprises optimize decision-making strategies, clarify the direction and focus of future decisions, and enhance the overall benefits and competitiveness of enterprises.

[0047] Preferably, step S4 comprises the following steps:

[0048] Step S41: constructing an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data to generate an intelligent decision tree model;

[0049] Step S42: Designing enterprise intelligent decision logic parameters based on the intelligent decision tree model to generate enterprise intelligent decision logic parameters;

[0050] Step S43: designing a preliminary intelligent business practice engine according to the enterprise intelligent decision-making logic parameters;

[0051] Step S44: Analyze the decision warning indicators of the preliminary intelligent business practice engine through the preset intelligent decision warning threshold, and design the engine decision warning mechanism according to the decision warning indicators;

[0052] Step S45: Design a supply business smart contract for the preliminary smart business practice engine using blockchain technology, generate a supply business smart contract, and design an engine supply business monitoring mechanism based on the supply business smart contract;

[0053] Step S46: Perform engine mechanism configuration processing on the preliminary intelligent business practice engine according to the engine decision warning mechanism and the engine supply business monitoring mechanism to generate an intelligent business practice engine.

[0054] The present invention constructs an intelligent decision tree model and uses historical enterprise decision feature data to perform intelligent decision analysis. The generated model can simulate and predict the effect of enterprise decision-making, improve the scientificity and accuracy of decision-making, and the intelligent decision tree model is trained and optimized based on data, which can make full use of the information in historical data and provide more accurate and objective decision support. The optimized decision logic parameters are designed through the intelligent decision tree model, and the enterprise intelligent decision logic parameters are generated, so that the enterprise decision process is more reasonable and efficient. The design of the intelligent decision logic parameters helps enterprises to automatically make higher-quality decisions in a complex business environment and improve the overall benefits of enterprise operations. According to the enterprise intelligent decision logic parameters, a preliminary intelligent business practice engine is designed, which can realize the intelligence of the business processing process, improve the automation and efficiency of business processing, and the intelligent business practice engine can be adjusted and optimized according to the actual business needs of the enterprise, and has strong flexibility and adaptability. By analyzing the decision warning indicators of the preliminary intelligent business practice engine, potential business risks and problems can be identified in advance, and a corresponding decision warning mechanism can be designed to help enterprises respond to and handle emergencies in a timely manner during business operation, and improve the enterprise's resilience and risk management level. By using blockchain technology to design supply business smart contracts, we can achieve transparency and credibility in the business transaction process, improve the security and reliability of business operations, and automatically execute preset business rules to reduce human intervention and improve the efficiency and accuracy of business processing. Through engine mechanism configuration processing, we can comprehensively optimize the functions and mechanisms of the intelligent business practice engine and generate a more complete and efficient intelligent business practice engine. The resulting intelligent business practice engine can provide strong support in the actual operation of the enterprise and improve the overall efficiency and intelligence level of business processing.

[0055] Preferably, step S41 includes the following steps:

[0056] Step S411: using a preset Bayesian network algorithm to perform decision feature probability distribution analysis on historical enterprise decision feature data to generate decision feature probability distribution data;

[0057] Step S412: construct an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data and decision feature probability distribution data to generate an intelligent decision tree model.

[0058] The present invention uses the Bayesian network algorithm to perform decision feature probability distribution analysis, accurately calculates the probability distribution of each decision feature, generates decision feature probability distribution data, provides more accurate probability information for the decision tree model, and improves the prediction ability of the model. The Bayesian network algorithm is good at handling uncertainty. Through probability distribution analysis, it better understands the dependencies between various decision features, and provides an important basis for the subsequent construction of decision models. Combining historical enterprise decision feature data and decision feature probability distribution data, a more optimized and accurate intelligent decision tree model is constructed to improve the guiding role of the model in enterprise decision-making. It uses rich historical data and probability distribution information, can more accurately simulate the actual decision-making scenarios of enterprises, and improve the accuracy and reliability of decisions.

[0059] Preferably, step S412 includes the following steps:

[0060] Perform enterprise decision splitting node analysis based on historical enterprise decision feature data to generate enterprise decision splitting nodes;

[0061] Conduct node splitting frequency analysis on enterprise decision splitting nodes to generate node splitting frequency data;

[0062] According to the node splitting frequency data, the splitting discrete degree analysis of the enterprise decision splitting node is carried out to generate the splitting discrete degree data;

[0063] Design the splitting node weight parameters according to the splitting discrete degree data and generate the splitting node weight parameters;

[0064] Performing decision tree model parameter analysis for enterprise decision analysis based on enterprise decision splitting nodes and splitting node weight parameters to obtain decision tree model parameters;

[0065] According to the probability distribution data of decision features, the decision tree model parameters are trained for enterprise decision intelligence to generate an intelligent decision tree model.

[0066] The present invention analyzes historical enterprise decision feature data, can accurately identify the split nodes in enterprise decision-making, generate enterprise decision split nodes, ensure the scientific nature of node selection in the process of building a decision tree model, and split node analysis helps to optimize the decision path, so that the decision tree model can better reflect the actual decision-making process of the enterprise and improve the applicability and accuracy of the model. Perform frequency analysis on the split nodes, identify the key nodes with higher frequency in the decision-making process, generate node split frequency data, help optimize the decision model structure, and the identification and analysis of high-frequency nodes help improve the construction efficiency of the decision tree model, ensuring that the model can respond and process quickly in the decision-making process. Split discrete degree analysis, understand the degree of dispersion of node distribution, generate split discrete degree data, help optimize the structure and hierarchy of the decision tree model, split discrete degree analysis helps to optimize the decision path selection, so that the model has better adaptability and flexibility when facing different decision-making scenarios. Through the split node weight parameter design, appropriate weights are assigned to each node, and split node weight parameters are generated to ensure that the model can reasonably consider the importance of each node in the decision-making process. The design of weight parameters helps to improve the accuracy of the decision tree model, so that the model can more accurately reflect the actual decision logic of the enterprise. By comprehensively considering the split nodes and weight parameters and analyzing the decision tree model parameters, the optimal model parameters can be obtained to ensure the scientificity and accuracy of the decision tree model. The optimized model parameters can significantly improve the performance of the decision tree model, making it more efficient and effective in practical applications. Using the decision feature probability distribution data to intelligently train the model can generate a more intelligent and accurate decision tree model, improve the model's prediction and decision-making capabilities, and the decision tree model after intelligent training can better simulate the actual decision-making scenarios of the enterprise, improve the quality of decision-making, and help enterprises make more scientific and effective decisions in complex business environments.

[0067] The beneficial effect of the present application is that the present invention ensures the scientificity and accuracy of the design of the system function module by comprehensively acquiring and finely analyzing the ERP demand data, and the correlation analysis and aggregation processing of the demand data effectively avoids the omission and duplication of requirements, improves the efficiency and accuracy of demand management, and enables the ERP system to better adapt to the actual business needs of the enterprise, and ensures the reasonable configuration and efficient operation of the function module. The microservice architecture mapping and containerization processing of the ERP function module improve the flexibility and scalability of the system. The microservice architecture enables the system to respond to changes in business needs more flexibly, improves the reliability and maintenance efficiency of the system, and the containerization processing ensures the consistency and stability of the system in different environments, simplifies the deployment and operation and maintenance process, and improves the speed and quality of the system going online. The in-depth analysis and processing of historical enterprise source data optimizes data quality and decision-making benefits. The enterprise decision gain based on feature data affects the benefit analysis, provides scientific decision support, improves the accuracy and effectiveness of enterprise decision-making, and helps enterprises make more scientific and efficient decisions in complex business environments. Through the design of intelligent decision tree model and intelligent business practice engine, the intelligence and automation of business processing are realized. The intelligent decision tree model uses historical data and probability distribution information to build an efficient decision model, which improves the accuracy of decision-making. The intelligent business practice engine improves the automation level and adaptability of business processing through intelligent business logic and early warning mechanism, ensuring that enterprises can quickly respond to market changes and business needs. Thus, a well-designed ERP system helps enterprises to carry out overall planning and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of the steps of an ERP system online method of the present invention;

[0069] Figure 2 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0070] Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0071] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0072] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0073] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0074] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0075] To achieve this, please refer to Figures 1 to 3 The present invention provides an ERP system online method, comprising the following steps:

[0076] Step S1: Acquire ERP demand data; perform ERP related demand analysis based on the ERP demand data to generate ERP related demand data; perform related demand data aggregation processing on the ERP related demand data to generate ERP related aggregated demand data;

[0077] Step S2: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate ERP-related function modules; ERP system architecture analysis is performed based on the ERP-related function modules to establish the ERP system architecture;

[0078] Step S3: Acquire historical enterprise source data; remove redundant features from the historical enterprise source data to generate historical enterprise feature data; perform historical enterprise decision feature analysis based on the historical enterprise feature data to generate historical enterprise decision feature data;

[0079] Step S4: Analyze the enterprise intelligent decision logic parameters according to the historical enterprise decision feature data to generate the enterprise intelligent decision logic parameters; design the intelligent business practice engine based on the enterprise intelligent decision logic parameters to generate the intelligent business practice engine;

[0080] Step S5: Based on the intelligent business practice engine, the intelligent business practice function configuration processing of the ERP system architecture is performed to generate an ERP system, and the ERP system is transmitted to the cloud server to execute the ERP system online operation.

[0081] The present invention can fully understand the specific needs of enterprises in various business modules by acquiring ERP demand data, provide basic data for subsequent demand analysis, and perform related demand analysis on ERP demand data, so as to accurately identify the related needs between various business modules and avoid omission and duplication of needs. Through the aggregation processing of related demand data, the scattered needs can be integrated, and comprehensive ERP related aggregation demand data can be generated, so as to improve the efficiency and accuracy of demand management. Functional module analysis is performed on ERP related aggregation demand data to generate functional modules for enterprise needs, improve the modularization degree of system design, and facilitate system maintenance and expansion. Through ERP system architecture analysis, the architecture of the system is reasonably planned to ensure that each module of the system can work efficiently and collaboratively, and improve the overall performance and stability of the system. Through redundant feature elimination of historical enterprise source data, useless or repeated data features can be removed, streamlined historical enterprise feature data can be generated, and data quality can be improved. Based on historical enterprise feature data, decision feature analysis is performed, data features valuable to enterprise decision-making can be excavated, historical enterprise decision feature data can be generated, and support can be provided for enterprise intelligent decision-making. Through enterprise intelligent decision logic parameter analysis, accurate decision logic parameters are generated, and scientific basis is provided for enterprise intelligent decision-making, and the accuracy and effectiveness of decision-making are improved. Designing an intelligent business practice engine based on intelligent decision-making logic parameters can integrate decision logic into business practice, form an intelligent business processing engine, and improve the intelligence level of enterprise business processing. Functional configuration processing of the system architecture by the intelligent business practice engine can realize intelligent configuration of system functions, ensuring that the ERP system can flexibly respond to the actual business needs of the enterprise. The generated ERP system is transmitted to the cloud server to perform the online operation, which can utilize the efficiency and flexibility of cloud computing to achieve rapid deployment and online of the ERP system, reducing the implementation cycle and cost. Therefore, an online method of an ERP system of the present invention ensures efficient coordination and integration between system modules through the association and aggregation analysis of functional modules, ensures the effective execution of the functions of the ERP system, and satisfies the user's intelligent experience service through the intelligent business practice engine design.

[0082] As an embodiment of the present invention, refer to Figure 1 The above is a schematic flow chart of the steps of an ERP system online method of the present invention. In this embodiment, the ERP system online method includes the following steps:

[0083] Step S1: Acquire ERP demand data; perform ERP related demand analysis based on the ERP demand data to generate ERP related demand data; perform related demand data aggregation processing on the ERP related demand data to generate ERP related aggregated demand data;

[0084] In an embodiment of the present invention, the ERP demand data is the demand information for the construction of an enterprise resource planning system (ERP system) that needs to be established. The ERP demand data is collected through interviews with various business departments, online demand information collection, etc., and these data cover the needs of finance, human resources, supply chain, production and sales. Data analysis software (such as SAP Analytics Cloud or self-developed tools) is used to identify and analyze the key demand points in the data and the dependencies between the demands. From the collected demand data, the association rule mining technology is used to identify the dependencies and interactions in the business process to generate ERP associated demand data. The associated demand data is classified and aggregated by functional modules (such as procurement, inventory, sales, etc.) through data aggregation tools to simplify the complexity of the demand data. The aggregate query function of the database management system (such as Microsoft SQL Server) is used to integrate the scattered demand data to generate ERP associated aggregated demand data, which provides preparation for subsequent module analysis.

[0085] Step S2: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate ERP-related function modules; ERP system architecture analysis is performed based on the ERP-related function modules to establish the ERP system architecture;

[0086] In the embodiment of the present invention, based on the ERP associated aggregation demand data, modular design principles and tools (such as UML modeling tools) are used to analyze and define the core functional modules of the ERP system, determine the necessary functional modules of the ERP system, such as financial management, inventory control, etc., and clarify the data and control flow relationships between the modules. Use architecture design software (such as microservice architecture technology and containerization technology) to design the system architecture for the associated functional modules, and design the architecture of the ERP system according to the relationship between the functional modules and the enterprise's IT strategy, including determining the system's physical deployment, data flow, interface, and integration solutions with other systems.

[0087] Step S3: Acquire historical enterprise source data; remove redundant features from the historical enterprise source data to generate historical enterprise feature data; perform historical enterprise decision feature analysis based on the historical enterprise feature data to generate historical enterprise decision feature data;

[0088] In the embodiment of the present invention, the internal database system of the enterprise, such as the ERP system, CRM system and other business systems, is used to collect historical enterprise data, including financial records, sales data, supply chain information, etc., and data integration tools (such as ETL tools) are used to ensure that data is extracted and integrated from different sources and data sources in different formats. Data cleaning software is applied or custom scripts are written (for example, using data processing libraries in Python or R) to identify and remove duplicate, irrelevant or missing data fields, retain features with analytical value, such as turnover, cost, inventory levels, etc., and ensure the accuracy and relevance of the data set. Statistical analysis software or machine learning algorithms (such as decision tree analysis and cluster analysis) are used to analyze the processed data, identify key decision factors, and determine feature data that has a significant impact on enterprise decisions, such as the most favorable sales channels, the most cost-effective production strategies, etc.

[0089] Step S4: Analyze the enterprise intelligent decision logic parameters according to the historical enterprise decision feature data to generate the enterprise intelligent decision logic parameters; design the intelligent business practice engine based on the enterprise intelligent decision logic parameters to generate the intelligent business practice engine;

[0090] In the embodiment of the present invention, based on the historical enterprise decision feature data, a machine learning algorithm is used to analyze and identify decision patterns and logic parameters, and decision logic parameters such as decision thresholds and key influencing factor weights are extracted from the analysis results to support intelligent decision making. Software development tools (such as Microsoft Visual Studio) and model building platforms (such as TensorFlow or PyTorch) are used to automate business practices, and an intelligent business practice engine is designed based on decision logic parameters to ensure that the engine can automatically execute key business processes, such as automated inventory management, intelligent price setting, optimized resource allocation and other decision features. After the design is completed, the engine is integrated into the enterprise ERP system to achieve real-time data processing and decision support.

[0091] Step S5: Based on the intelligent business practice engine, the intelligent business practice function configuration processing of the ERP system architecture is performed to generate an ERP system, and the ERP system is transmitted to the cloud server to execute the ERP system online operation.

[0092] In the embodiment of the present invention, an intelligent business practice engine is configured in the ERP system architecture. This includes setting parameters, rules and logic related to business processes to ensure that the engine can automatically make business decisions based on real-time data. After the intelligent business practice engine is integrated in the system architecture, a comprehensive integration test is performed. The purpose of the test is to verify whether the engine's functions meet business requirements and ensure that data flows can flow correctly between the various modules of the ERP system. Using the ERP system development kit (SDK) and application programming interface (APIs), based on the configured intelligent business practice functions, build various modules of the ERP system, such as finance, human resources, supply chain management, etc., design and implement the user interface of the ERP system, ensure that the interface is friendly, and can effectively display necessary business information and analysis results, and support users in daily operations and decision-making. According to the needs of the enterprise, select a suitable cloud platform (such as Amazon Web Services, Microsoft Azure, Google Cloud Platform, etc.) to deploy the ERP system, configure the required cloud resources, including servers, storage and network resources, and ensure that the performance and availability of the system meet the needs of the enterprise. Deploy the ERP system to a configured cloud server, including installing the system, configuring the database, setting up network connections, etc. Perform a final test of the system in a cloud environment to check the overall functionality, performance, and security of the system, and ensure that all components are ready and can operate normally. After all tests are completed, the ERP system is officially launched to serve the company's daily operations.

[0093] Preferably, step S1 comprises the following steps:

[0094] Step S11: Obtain ERP demand data;

[0095] Step S12: using natural language technology to perform ERP demand information mining on standard ERP demand data to generate ERP demand mining data;

[0096] Step S13: performing ERP-related demand analysis on the ERP demand mining data to generate ERP-related demand data;

[0097] Step S14: performing associated demand data aggregation processing on the ERP associated demand data to generate ERP associated aggregated demand data.

[0098] The present invention can comprehensively collect the specific needs of each business module of an enterprise by acquiring ERP demand data, provide basic data for subsequent analysis and processing, and avoid misjudgment of needs due to incomplete or inaccurate data. By using natural language processing technology, valuable information can be efficiently extracted from standard ERP demand data to generate structured ERP demand mining data. Natural language processing technology can understand and parse complex text information, improve the accuracy of demand mining, and avoid errors and omissions in manual analysis. Through the association analysis of ERP demand mining data, the associated needs between different business modules can be identified, the isolated design of each module demand can be avoided, and each demand can be prioritized according to the associated demand data to ensure that key needs are given priority, thereby improving the scientificity and rationality of system design. Through aggregation processing, the scattered associated demand data are integrated to generate comprehensive and comprehensive ERP associated aggregated demand data, which is convenient for the design of subsequent functional modules and the construction of system architecture. Aggregation processing can effectively reduce redundancy and conflict in demand data, ensure the consistency and integrity of demand data, and improve the efficiency and quality of system design.

[0099] In the embodiment of the present invention, ERP system requirements are collected through the internal online information integration system of the enterprise, and these requirements data are recorded and classified using special requirements management software (such as JIRA or Trello), ensuring that all collected information is properly recorded and tracked to obtain ERP requirements data. Natural language processing technology (NLP), especially text mining and semantic analysis tools (such as IBM Watson or Google Cloud Natural Language API), is applied to analyze and process ERP requirements text data, and these technologies are used to identify keywords and phrases, extract important information in the requirements, such as functional descriptions, performance requirements, etc., and convert them into structured ERP requirements mining data. Data analysis software (such as Tableau or Microsoft Power BI) is used to analyze ERP requirements mining data, and by pre-establishing a requirements association model, the dependencies and potential connections between different requirements are identified, and the ERP associated requirements data is extracted from the association and dependency information obtained in the analysis, providing a basis for further aggregation processing and system design. Through data aggregation tools (such as the aggregate query function of the SQL database), ERP-related demand data can be aggregated by function, department or any other business logic, including merging similar demands, unifying the demand expression format, etc. The processed data will be more systematic and simplified, and easy to apply and implement in the subsequent ERP system design, generating ERP-related aggregated demand data.

[0100] Preferably, step S13 comprises the following steps:

[0101] Step S131: Analyze and process ERP demand entity information according to ERP demand mining data to generate ERP demand entity data;

[0102] Step S132: performing ERP demand dependency analysis according to the ERP demand entity data to generate ERP demand dependency data;

[0103] Step S133: using the word frequency-inverse document frequency technology to convert the ERP demand mining data into numerical vectors to generate ERP demand mining vector data;

[0104] Step S134: performing ERP demand similarity analysis based on the ERP demand mining vector data to generate ERP demand similarity data;

[0105] Step S135: performing ERP-related demand analysis on the ERP demand mining data based on the ERP demand dependency data and the ERP demand similarity data to generate ERP-related demand data.

[0106] The present invention can clearly identify each specific demand object through ERP demand entity information analysis, generate ERP demand entity data, improve the accuracy of demand analysis, convert complex demand text data into structured entity data, facilitate subsequent processing and analysis, and improve the efficiency and accuracy of data processing. By analyzing the dependency relationship between ERP demand entity data, the mutual relationship between each demand can be identified, and ERP demand dependency relationship data can be generated, which is helpful for the overall design of the system, clarifies the dependency relationship between demands, facilitates the priority sorting of demands and the reasonable allocation of resources, and improves the scientificity and rationality of demand processing. Through the term frequency-inverse document frequency (TF-IDF) technology, the text data is converted into numerical vector data, the representation form of the demand data is standardized, and the uniformity of data processing is improved. The TF-IDF technology can effectively extract key information in the demand text and improve the accuracy and effectiveness of data analysis. Through similarity analysis, similar parts in the demand data can be identified, and ERP demand similarity data can be generated. Similarity analysis helps to perform association analysis on similar demands, facilitates the integration and optimization of demands, and improves the scientificity of demand management. Comprehensively consider the dependency and similarity data of requirements, conduct a comprehensive correlation analysis on the demand mining data, generate ERP related demand data, ensure the comprehensiveness and accuracy of the demand analysis, and through the correlation demand analysis, provide more accurate demand data support for system design and improve the scientificity and rationality of the system functional module design.

[0107] In an embodiment of the present invention, important entity information, such as functional requirements, performance requirements, user roles, etc., is extracted from ERP demand text mining data, and entity recognition tools (such as Spacy or Stanford NER) are used to automatically identify and extract these entity information, ensuring that the key elements in each demand are accurately identified, and the identified entity information is structured and stored in the database to form ERP demand entity data, which is convenient for subsequent analysis and processing. A dependency model between demand entities is established using a graph database (such as Neo4j). Each entity is a node in the graph, and the dependency relationship is used as an edge to connect each node. Through the query and analysis function of the graph database, the dependency relationship and hierarchical structure between each demand entity are identified. According to the analysis results, ERP demand dependency relationship data is generated and stored in the system for further use. The term frequency-inverse document frequency (TF-IDF) technology is applied to convert the ERP demand text data into a numerical vector, and the machine learning library such as Scikit-learn is used for TF-IDF processing. Each demand text is converted into a vector, the dimension of the vector represents different terms, and the numerical value represents the importance of the term in the document. The converted vector data is stored in the database to form ERP demand mining vector data. Use the cosine similarity algorithm to calculate the similarity between each pair of ERP demand vectors, use the SciPy library in Python or similar tools for calculation, identify similar demand pairs based on the calculation results, and quantify the similarity between each pair of demands, store these similarity information in a structured manner, and generate ERP demand similarity data to facilitate subsequent aggregation and optimization. Integrate ERP demand dependency data and similarity data to build a comprehensive demand association model, use association analysis tools (such as association rule mining algorithm Apriori) to analyze the integrated data, identify the complex associations between the demands, store the analysis results in a structured manner, and generate ERP associated demand data to provide a basis for system design and optimization.

[0108] Preferably, step S2 comprises the following steps:

[0109] Step S21: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate an ERP-related function module;

[0110] Step S22: performing microservice architecture mapping processing of the functional modules according to the ERP-related functional modules to generate microservice architecture mapping data;

[0111] Step S23: Containerize the microservice architecture mapping data to generate containerized microservice architecture mapping data, and establish an ERP system architecture through the containerized microservice architecture mapping data.

[0112] The present invention analyzes the functional modules of ERP associated aggregated demand data, and can reasonably divide the demand data into multiple functional modules, improve the modularity of system design, and the division of functional modules helps to clarify the division of labor, improve the efficiency of system design and development, and facilitate the independent development and testing of modules. Mapping the functional modules to the microservice architecture can improve the flexibility and scalability of the system, so that the system can quickly respond to changes in business needs. The microservice architecture can split the system functions into multiple independent services, improve the reliability and stability of the system, and reduce the impact of single point failures. Through containerization, the consistency of each microservice in different environments can be ensured, the problems caused by environmental differences can be reduced, and the stability of system deployment can be improved. The containerization technology makes the deployment and operation and maintenance of the system simpler and more efficient, and can achieve rapid deployment, expansion and migration, improve the efficiency of system launch and maintenance, provide good resource isolation and management capabilities, and can better utilize system resources and improve the performance and security of the system.

[0113] In an embodiment of the present invention, a demand analysis tool (such as a UML modeling tool or other system design software) is used to map ERP-related aggregated demand data to specific functional modules. For example, the demand data includes multiple modules such as financial management, inventory management, and human resource management. Based on the relevance and functional similarity of the demand data, the specific functions and operation processes of each functional module are defined. Ensure that each module operates independently but can cooperate with each other, store the designed functional modules in a structured manner, and generate ERP-related functional modules to provide a basis for system architecture design. Use a microservice architecture design tool (such as SpringBoot or Docker Compose) to map ERP-related functional modules to a microservice architecture, each functional module corresponds to one or more microservices, ensure loose coupling and high cohesion between modules, define interfaces and communication protocols (such as RESTful API) for each microservice, ensure data exchange and function calls between services, and use API design tools (such as Swagger) to standardize interface design. Map the data requirements and processing logic of the ERP-related functional modules to the internal data model of each microservice. Ensure that each microservice can independently process its corresponding data operations, digitize the microservice architecture design and interface definition, generate microservice architecture mapping data, and provide a basis for subsequent containerization processing. Use containerization technology tools such as Docker or Kubernetes to containerize microservice mapping data. Each microservice runs in an independent container to ensure environmental isolation and deployment consistency. Write a Dockerfile for each microservice to define the container build process, including basic images, dependency installation, code deployment, and startup commands. Use CI / CD tools (such as Jenkins or GitLab CI) to automatically build and deploy containers. Use container orchestration tools such as Kubernetes to manage and deploy multiple container instances, and define strategies such as service discovery, load balancing, and automatic expansion. Write Kubernetes configuration files (such as YAML files) to define the deployment, service, configuration, and persistent storage of each service in the cluster. Through containerization, deploy all microservices to the container orchestration platform to establish a complete ERP system architecture to ensure that the system architecture is highly available, scalable, and easy to maintain.

[0114] Preferably, step S3 comprises the following steps:

[0115] Step S31: Obtain historical enterprise source data;

[0116] Step S32: performing cluster analysis on the historical enterprise source data based on a preset clustering algorithm to generate historical enterprise cluster data;

[0117] Step S33: Calculate the clustering silhouette coefficient based on the historical enterprise clustering data to generate the clustering silhouette coefficient;

[0118] Step S34: performing redundant information analysis on the historical enterprise source data according to the clustering silhouette coefficient to generate historical enterprise redundant data;

[0119] Step S35: Eliminate redundant features of historical enterprise source data according to historical enterprise redundant data to generate historical enterprise feature data;

[0120] Step S36: Performing an enterprise decision gain impact benefit analysis based on historical enterprise characteristic data to generate enterprise decision gain impact benefit data;

[0121] Step S37: extracting enterprise decision features of the enterprise decision gain affecting benefits from the historical enterprise feature data according to the enterprise decision gain affecting benefits data, and generating historical enterprise decision feature data.

[0122] The present invention can provide a solid foundation for subsequent data analysis and processing by acquiring historical enterprise source data, ensure the comprehensiveness and accuracy of data, collect and integrate enterprise historical data, help summarize and analyze the past operation and decision-making of enterprises, and provide reference for future decision-making. Through cluster analysis, historical enterprise source data are reasonably grouped to generate historical enterprise cluster data, which is helpful to identify patterns and trends in data. Cluster analysis can effectively reduce the complexity of data and improve the efficiency of data processing and analysis. By calculating the clustering silhouette coefficient, the effect of cluster analysis is evaluated to ensure the rationality and accuracy of clustering results. The clustering silhouette coefficient helps to identify and optimize clustering algorithm parameters and improve the quality of cluster analysis. Through redundant information analysis, redundant information in historical enterprise source data can be identified and eliminated, historical enterprise redundant data is generated, data structure is optimized, data redundancy is reduced, data quality and consistency are improved, and higher quality data support is provided for subsequent data processing and analysis. The elimination of redundant features generates streamlined historical enterprise feature data, reduces the complexity of data processing and storage costs, and after removing redundant features, the data analysis results are more accurate and the decision-making basis is more reliable. The analysis of the gain-effect benefit of historical enterprise characteristic data can evaluate the benefits of enterprise decisions, generate enterprise decision-making gain-effect benefit data, provide scientific decision-making basis for enterprises, help enterprises identify high-efficiency decision-making models, optimize the decision-making process, and improve the overall benefits of enterprises. The analysis of the gain-effect benefit data of enterprise decisions can extract the key features that have an important impact on enterprise decisions, generate historical enterprise decision-making feature data, improve the scientificity and effectiveness of enterprises in future decisions, and enhance the competitiveness of enterprises.

[0123] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 FIG. 1 is a flow chart of detailed implementation steps of step S3 in FIG. 1 . In this embodiment, step S3 includes:

[0124] Step S31: Obtain historical enterprise source data;

[0125] In an embodiment of the present invention, historical data including financial records, sales data, inventory information, supply chain data, etc. are extracted from an existing information system of an enterprise (such as an existing ERP system, CRM system, financial system, etc.), and ETL (Extract, Transform, Load) tools (such as Talend, Informatica or Apache NiFi) are used to extract and integrate these data, and the extracted data is stored in a unified database or data warehouse to ensure the integrity and consistency of the data.

[0126] Step S32: performing cluster analysis on the historical enterprise source data based on a preset clustering algorithm to generate historical enterprise cluster data;

[0127] In the embodiment of the present invention, an appropriate clustering algorithm (such as K-means, DBSCAN or hierarchical clustering) is selected, and the specific selection depends on the characteristics of the data and the analysis objectives. The data is preprocessed, including standardization, missing value filling and feature selection, to ensure that the data is suitable for cluster analysis. The preprocessed data is analyzed using a clustering algorithm to identify patterns and groups in the data and generate historical enterprise clustering data. Cluster analysis is implemented using data analysis software (such as Python's Scikit-learn or R's cluster package).

[0128] Step S33: Calculate the clustering silhouette coefficient based on the historical enterprise clustering data to generate the clustering silhouette coefficient;

[0129] In the embodiment of the present invention, the silhouette coefficient of each data point is calculated to measure its matching degree in the cluster. The value range of the silhouette coefficient is [-1, 1]. The larger the value, the better the clustering effect. The cluster silhouette coefficient is calculated using a data analysis tool (such as the silhouette_score function in Python's Scikit-learn), and the calculated cluster silhouette coefficient is stored in a database to generate cluster silhouette coefficient data for subsequent analysis.

[0130] Step S34: performing redundant information analysis on the historical enterprise source data according to the clustering silhouette coefficient to generate historical enterprise redundant data;

[0131] In an embodiment of the present invention, redundant information and outliers in the data are identified based on the clustering silhouette coefficient. Data points with low clustering silhouette coefficients are redundant or outliers. Data analysis tools (such as Python's Pandas or R's dplyr package) are used to screen and analyze the clustering data to identify redundant information, and the identified redundant data is structured and stored to generate historical enterprise redundant data, providing a basis for subsequent feature elimination.

[0132] Step S35: Eliminate redundant features of historical enterprise source data according to historical enterprise redundant data to generate historical enterprise feature data;

[0133] In the embodiment of the present invention, unnecessary features and data points are removed based on the analysis results of redundant data. It is ensured that the data set only retains features that are useful for decision-making, and redundant features are removed using data cleaning tools (such as OpenRefine and Python's Pandas library), and the processed data is stored as historical enterprise feature data to ensure that the data set is concise and has a high amount of information.

[0134] Step S36: Performing an enterprise decision gain impact benefit analysis based on historical enterprise characteristic data to generate enterprise decision gain impact benefit data;

[0135] In the embodiment of the present invention, regression analysis of machine learning algorithms and other methods are used to evaluate the impact of different decisions on corporate benefits, and data analysis and modeling tools (such as Python's Statsmodels and R's caret package) are used to perform benefit analysis. Based on the analysis results, corporate decision-making gain impact benefit data is generated.

[0136] Step S37: extracting enterprise decision features of the enterprise decision gain affecting benefits from the historical enterprise feature data according to the enterprise decision gain affecting benefits data, and generating historical enterprise decision feature data.

[0137] In an embodiment of the present invention, based on the benefit analysis results, features that have a significant impact on decision gains are extracted, feature selection algorithms (such as recursive feature elimination and principal component analysis) are used to perform feature extraction, and feature extraction tools (such as Python's Scikit-learn and R's caret package) are used to implement feature extraction. The extracted key decision features are stored as historical enterprise decision feature data to ensure that these features can be used for subsequent intelligent decision logic analysis.

[0138] Preferably, step S32 includes the following steps:

[0139] Step S321: Calculate the intra-cluster distance based on the historical enterprise clustering data to generate the intra-cluster distance data;

[0140] Step S322: Calculate the inter-cluster distance based on the historical enterprise clustering data to generate inter-cluster distance data;

[0141] Step S323: Calculate the clustering silhouette coefficient based on the intra-cluster distance data and the inter-cluster distance data to generate the clustering silhouette coefficient.

[0142] The present invention calculates the distance within the cluster, evaluates the closeness of the data points within the same cluster, generates the distance data within the cluster, helps to understand the internal consistency of each cluster, and the calculation result of the distance within the cluster helps to optimize the clustering algorithm, ensures that the data points within the same cluster are closer, thereby improving the accuracy of cluster analysis. By calculating the distance between clusters, the independence between different clusters is evaluated, and the distance data between clusters is generated to ensure the separation and non-overlap of the clustering results. The calculation result of the distance between clusters can be used to improve the clustering algorithm, so that the distance between different clusters is larger, and the effect and accuracy of cluster analysis are improved. Through the calculation of the clustering silhouette coefficient, the quality of the clustering results can be comprehensively evaluated, the clustering silhouette coefficient is generated, and the quality of the clustering effect is understood as a whole. The clustering silhouette coefficient provides a basis for the adjustment of the clustering algorithm, helps to identify problems in the clustering results, and performs corresponding optimization and improvement to improve the quality and reliability of cluster analysis.

[0143] In the embodiment of the present invention, the data points of each cluster are extracted based on the historical enterprise cluster data. The intra-cluster distance refers to the distance between the data points in the same cluster. These distances are the distances from each data point to the cluster center (centroid). The distance between each data point and its cluster center is calculated, and the average value or the average value between all data points is taken to determine the center point (centroid) of each cluster. These center points represent the average position of each cluster. The distance from each data point to the center of the cluster to which it belongs is calculated. Finally, the average value of all these distances is taken, and the calculated intra-cluster distance of each cluster is stored as structured data to generate cluster intra-cluster distance data. The inter-cluster distance refers to the distance between different clusters. These distances refer to the distance between the center points of each cluster. The distance between each pair of cluster center points is calculated by Euclidean distance, which represents the degree of separation between different clusters. The calculated inter-cluster distance between each pair of clusters is stored as structured data to facilitate subsequent analysis and use, and generate cluster inter-cluster distance data. The silhouette coefficient combines the intra-cluster distance and the inter-cluster distance to measure the clustering quality of each data point. The silhouette coefficient of each data point is calculated using the intra-cluster distance data and the inter-cluster distance data. The silhouette coefficients of all data points are averaged to obtain the clustering silhouette coefficient.

[0144] Preferably, step S36 includes the following steps:

[0145] Step S361: Evaluate the impact of historical enterprise decisions on benefits based on historical enterprise characteristic data to generate historical enterprise decision impact benefit data;

[0146] Step S362: Perform enterprise decision gain impact benefit analysis based on historical enterprise decision impact benefit data to generate enterprise decision gain impact benefit data.

[0147] The present invention evaluates historical enterprise characteristic data, can understand the actual impact and benefits of past enterprise decisions, generate historical enterprise decision-making impact benefit data, and help enterprises summarize past decision-making experiences and lessons. The historical enterprise decision-making impact benefit data provides an important reference basis for future decisions, helping enterprises better understand which decisions have a positive impact on enterprise benefits and which decisions may have problems. The analysis of historical enterprise decision-making impact benefit data can identify the decisions that have the greatest impact on enterprise gains, generate enterprise decision-making gain impact benefit data, and help enterprises optimize the decision-making process and improve the scientificity and accuracy of decisions. Enterprise decision-making gain impact benefit data can help enterprises optimize decision-making strategies, clarify the direction and focus of future decisions, and enhance the overall benefits and competitiveness of enterprises.

[0148] In the embodiment of the present invention, statistical analysis and regression analysis are used to evaluate the benefits of historical decisions of historical enterprise feature data. For example, multiple regression analysis is used to determine the degree of influence of each feature on enterprise benefits. By establishing a regression model, the feature data is used as an independent variable, and the enterprise benefit data (such as profit, income) is used as a dependent variable to evaluate the contribution of each feature to the enterprise benefit. Data analysis tools (such as Python's Statsmodels and R's lm function) are used for regression analysis to quantify the impact of each feature. The stability and predictive ability of the model are evaluated through cross-validation (such as K-fold cross-validation) to ensure the reliability of the analysis results. According to the analysis results, historical enterprise decision-making impact benefit data are generated to record the contribution and influence of each feature on the enterprise benefit, providing a basis for subsequent decision analysis. Integrate historical enterprise decision-making impact benefit data into an analysis framework as the basic data for enterprise decision-making gain impact benefit analysis, apply gain analysis models (such as cost-benefit analysis model, gain matrix, etc.), evaluate the gain effect of different decision options on enterprise benefits, compare the benefits of different historical decisions under different conditions, identify decision-making patterns and strategies that can maximize enterprise benefits, evaluate the effects of decisions in different market and operating environments through scenario analysis and sensitivity analysis, ensure the comprehensiveness of analysis results, use data analysis tools (such as Tableau, Power BI) for visual analysis, display the gain effect and impact of different decision options, combine machine learning algorithms to further refine the analysis, and identify key factors in complex decision-making patterns. Based on the analysis results, generate enterprise decision-making gain impact benefit data, record the gain benefit indicators of each decision option, and form enterprise decision-making gain impact benefit data for decision guidance and optimization.

[0149] Preferably, step S4 comprises the following steps:

[0150] Step S41: constructing an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data to generate an intelligent decision tree model;

[0151] Step S42: Designing enterprise intelligent decision logic parameters based on the intelligent decision tree model to generate enterprise intelligent decision logic parameters;

[0152] Step S43: designing a preliminary intelligent business practice engine according to the enterprise intelligent decision-making logic parameters;

[0153] Step S44: Analyze the decision warning indicators of the preliminary intelligent business practice engine through the preset intelligent decision warning threshold, and design the engine decision warning mechanism according to the decision warning indicators;

[0154] Step S45: Design a supply business smart contract for the preliminary smart business practice engine using blockchain technology, generate a supply business smart contract, and design an engine supply business monitoring mechanism based on the supply business smart contract;

[0155] Step S46: Perform engine mechanism configuration processing on the preliminary intelligent business practice engine according to the engine decision warning mechanism and the engine supply business monitoring mechanism to generate an intelligent business practice engine.

[0156] The present invention constructs an intelligent decision tree model and uses historical enterprise decision feature data to perform intelligent decision analysis. The generated model can simulate and predict the effect of enterprise decision-making, improve the scientificity and accuracy of decision-making, and the intelligent decision tree model is trained and optimized based on data, which can make full use of the information in historical data and provide more accurate and objective decision support. The optimized decision logic parameters are designed through the intelligent decision tree model, and the enterprise intelligent decision logic parameters are generated, so that the enterprise decision process is more reasonable and efficient. The design of the intelligent decision logic parameters helps enterprises to automatically make higher-quality decisions in a complex business environment and improve the overall benefits of enterprise operations. According to the enterprise intelligent decision logic parameters, a preliminary intelligent business practice engine is designed, which can realize the intelligence of the business processing process, improve the automation and efficiency of business processing, and the intelligent business practice engine can be adjusted and optimized according to the actual business needs of the enterprise, and has strong flexibility and adaptability. By analyzing the decision warning indicators of the preliminary intelligent business practice engine, potential business risks and problems can be identified in advance, and a corresponding decision warning mechanism can be designed to help enterprises respond to and handle emergencies in a timely manner during business operation, and improve the enterprise's resilience and risk management level. By using blockchain technology to design supply business smart contracts, we can achieve transparency and credibility in the business transaction process, improve the security and reliability of business operations, and automatically execute preset business rules to reduce human intervention and improve the efficiency and accuracy of business processing. Through engine mechanism configuration processing, we can comprehensively optimize the functions and mechanisms of the intelligent business practice engine and generate a more complete and efficient intelligent business practice engine. The resulting intelligent business practice engine can provide strong support in the actual operation of the enterprise and improve the overall efficiency and intelligence level of business processing.

[0157] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed implementation step flow chart of step S4 in the embodiment, step S4 in this embodiment includes:

[0158] Step S41: constructing an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data to generate an intelligent decision tree model;

[0159] In an embodiment of the present invention, key feature data are extracted from historical enterprise decision feature data, and these data are used to train a decision tree model. A decision tree algorithm (such as CART, ID3 or C4.5) is selected, and the model is trained according to the enterprise's historical decision feature data. The decision tree model is constructed using the training data set. During the model training process, parameter adjustment and optimization are required to ensure the accuracy and stability of the model. Data analysis and machine learning tools (such as Python's Scikit-learn and R's rpart package) are used to construct and train the model, and cross-validation is performed to evaluate the generalization ability of the model to ensure that the model performs consistently on different data sets, so as to generate an intelligent decision tree model.

[0160] Step S42: Designing enterprise intelligent decision logic parameters based on the intelligent decision tree model to generate enterprise intelligent decision logic parameters;

[0161] In an embodiment of the present invention, important decision logic parameters are extracted from the intelligent decision tree model. These parameters include node splitting criteria, decision rules for leaf nodes, etc., and each node in the decision tree model is analyzed to determine the importance of each decision path and node. Decision logic parameters, such as thresholds and weights, are defined to ensure that these parameters can effectively guide the intelligent decision-making of the enterprise. Data analysis tools (such as Tableau and Excel) are used to analyze and design the decision logic parameters, and the designed intelligent decision logic parameters are recorded and stored to facilitate the subsequent design of the intelligent business practice engine.

[0162] Step S43: designing a preliminary intelligent business practice engine according to the enterprise intelligent decision-making logic parameters;

[0163] In the embodiment of the present invention, the architecture and functional modules of the intelligent business practice engine are designed based on the enterprise intelligent decision-making logic parameters. The engine needs to include a decision module, a data processing module, an execution module, etc., define the interface and communication method between the modules, ensure that the engine can run efficiently, use software development tools (such as Java, Python, C#) to implement the modules of the engine, and perform unit testing and integration testing to ensure that the engine's functions and performance meet expectations. The developed preliminary intelligent business practice engine is deployed in a test environment, ready for further optimization and configuration.

[0164] Step S44: Analyze the decision warning indicators of the preliminary intelligent business practice engine through the preset intelligent decision warning threshold, and design the engine decision warning mechanism according to the decision warning indicators;

[0165] In the embodiment of the present invention, intelligent decision warning thresholds are pre-set based on historical data and enterprise business needs. These thresholds are used to monitor business indicators, identify potential risks, and use the intelligent business practice engine to analyze the intelligent decision warning thresholds of real-time business data to generate warning indicators. An automated engine warning mechanism is designed based on the warning indicators, and data analysis and monitoring tools (such as Python's Pandas and Matplotlib, and monitoring systems such as Prometheus) are used to perform warning judgments and warning mechanism design, define warning levels, trigger conditions, and response measures, and implement automatic warning functions. When business indicators exceed the warning threshold, the system automatically generates a warning notification.

[0166] Step S45: Design a supply business smart contract for the preliminary smart business practice engine using blockchain technology, generate a supply business smart contract, and design an engine supply business monitoring mechanism based on the supply business smart contract;

[0167] In the embodiment of the present invention, blockchain technology is used to design supply business smart contracts, define the business rules and execution logic of the contracts, and these contracts automatically collect information on transactions and operations in the supply chain. Select a suitable blockchain platform (such as Ethereum, Hyperledger) to deploy smart contracts, write smart contract code, define the business rules and terms of the contract, deploy the smart contract to the blockchain platform, and test and verify to ensure the accuracy and security of the contract. Design a monitoring mechanism to monitor the execution of supply chain business in real time, define monitoring indicators and monitoring methods, and ensure the transparency and efficient operation of supply chain business. Use blockchain development tools (such as Solidity, Hyperledger Composer) to design and implement smart contracts, deploy and test smart contracts, and ensure their normal operation on the blockchain.

[0168] Step S46: Perform engine mechanism configuration processing on the preliminary intelligent business practice engine according to the engine decision warning mechanism and the engine supply business monitoring mechanism to generate an intelligent business practice engine.

[0169] In the embodiment of the present invention, the preliminary intelligent business practice engine is configured according to the designed decision-making warning mechanism and supply business monitoring mechanism to ensure that the engine can automatically monitor and respond to business changes, perform performance optimization and function adjustment, and ensure the efficient operation of the engine in the actual business environment. Through simulation testing, the stability and reliability of the engine are evaluated, and necessary adjustments and improvements are made. After completing the configuration and optimization, the final intelligent business practice engine is deployed and comprehensively tested to ensure that its functions and performance meet expectations.

[0170] Preferably, step S41 includes the following steps:

[0171] Step S411: using a preset Bayesian network algorithm to perform decision feature probability distribution analysis on historical enterprise decision feature data to generate decision feature probability distribution data;

[0172] Step S412: construct an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data and decision feature probability distribution data to generate an intelligent decision tree model.

[0173] The present invention uses the Bayesian network algorithm to perform decision feature probability distribution analysis, accurately calculates the probability distribution of each decision feature, generates decision feature probability distribution data, provides more accurate probability information for the decision tree model, and improves the prediction ability of the model. The Bayesian network algorithm is good at handling uncertainty. Through probability distribution analysis, it better understands the dependencies between various decision features, and provides an important basis for the subsequent construction of decision models. Combining historical enterprise decision feature data and decision feature probability distribution data, a more optimized and accurate intelligent decision tree model is constructed to improve the guiding role of the model in enterprise decision-making. It uses rich historical data and probability distribution information, can more accurately simulate the actual decision-making scenarios of enterprises, and improve the accuracy and reliability of decisions.

[0174] In the embodiment of the present invention, a suitable Bayesian network algorithm (such as a structural learning algorithm and a parameter learning algorithm) is selected to analyze the historical enterprise decision feature data to establish a probability relationship between the features, perform data cleaning and preprocessing, including processing missing values, standardizing data and encoding categorical variables, and using a binning method to convert continuous variables into discrete variables to adapt to the requirements of the Bayesian network. A Bayesian network model is constructed, the dependency relationship and probability distribution between the features are defined, a structural learning algorithm (such as Hill Climbing and Tabu Search) is used to determine the topological structure of the Bayesian network, and a parameter learning algorithm (such as maximum likelihood estimation and Bayesian estimation) is applied to calculate the conditional probability distribution of each node. The probability distribution of each feature is calculated by inference of the Bayesian network model to generate decision feature probability distribution data. The historical enterprise decision feature data and the decision feature probability distribution data are integrated into a comprehensive data set to provide comprehensive data support for the construction of a decision tree model, and a suitable decision tree algorithm (such as CART, ID3, and C4.5) is selected to construct an intelligent decision tree model. According to the specific application scenario and data characteristics, the most suitable algorithm is selected. The decision tree model is trained using the integrated data set. During the training process, methods such as cross-validation and grid search are used to optimize model parameters, determine the model's node splitting criteria, pruning strategies, etc., to ensure the accuracy and generalization ability of the model, evaluate the performance of the model, and use evaluation indicators such as accuracy, precision, recall rate and F1 score to measure the performance of the model. The classification effect of the model is deeply analyzed through methods such as confusion matrix and ROC curve. The trained intelligent decision tree model is saved, and the model's structure and parameter settings are recorded to ensure that the model can explain the decision-making process and provide visual decision paths and node information to generate an intelligent decision tree model.

[0175] Preferably, step S412 includes the following steps:

[0176] Perform enterprise decision splitting node analysis based on historical enterprise decision feature data to generate enterprise decision splitting nodes;

[0177] Conduct node splitting frequency analysis on enterprise decision splitting nodes to generate node splitting frequency data;

[0178] According to the node splitting frequency data, the splitting discrete degree analysis of the enterprise decision splitting node is carried out to generate the splitting discrete degree data;

[0179] Design the splitting node weight parameters according to the splitting discrete degree data and generate the splitting node weight parameters;

[0180] Performing decision tree model parameter analysis for enterprise decision analysis based on enterprise decision splitting nodes and splitting node weight parameters to obtain decision tree model parameters;

[0181] According to the probability distribution data of decision features, the decision tree model parameters are trained for enterprise decision intelligence to generate an intelligent decision tree model.

[0182] The present invention analyzes historical enterprise decision feature data, can accurately identify the split nodes in enterprise decision-making, generate enterprise decision split nodes, ensure the scientific nature of node selection in the process of building a decision tree model, and split node analysis helps to optimize the decision path, so that the decision tree model can better reflect the actual decision-making process of the enterprise and improve the applicability and accuracy of the model. Perform frequency analysis on the split nodes, identify the key nodes with higher frequency in the decision-making process, generate node split frequency data, help optimize the decision model structure, and the identification and analysis of high-frequency nodes help improve the construction efficiency of the decision tree model, ensuring that the model can respond and process quickly in the decision-making process. Split discrete degree analysis, understand the degree of dispersion of node distribution, generate split discrete degree data, help optimize the structure and hierarchy of the decision tree model, split discrete degree analysis helps to optimize the decision path selection, so that the model has better adaptability and flexibility when facing different decision-making scenarios. Through the split node weight parameter design, appropriate weights are assigned to each node, and split node weight parameters are generated to ensure that the model can reasonably consider the importance of each node in the decision-making process. The design of weight parameters helps to improve the accuracy of the decision tree model, so that the model can more accurately reflect the actual decision logic of the enterprise. By comprehensively considering the split nodes and weight parameters and analyzing the decision tree model parameters, the optimal model parameters can be obtained to ensure the scientificity and accuracy of the decision tree model. The optimized model parameters can significantly improve the performance of the decision tree model, making it more efficient and effective in practical applications. Using the decision feature probability distribution data to intelligently train the model can generate a more intelligent and accurate decision tree model, improve the model's prediction and decision-making capabilities, and the decision tree model after intelligent training can better simulate the actual decision-making scenarios of the enterprise, improve the quality of decision-making, and help enterprises make more scientific and effective decisions in complex business environments.

[0183] In an embodiment of the present invention, each feature and corresponding decision result are extracted from the historical enterprise decision feature data. Ensure that the data set contains comprehensive decision information and feature variables, use a decision tree algorithm (such as CART, ID3, C4.5), and generate a decision tree by recursively splitting the data set. At each splitting node, select the best splitting feature and splitting point according to the feature selection criteria (such as information gain, Gini coefficient), record each splitting node and its splitting conditions, and generate an enterprise decision splitting node. Count the frequency of occurrence of each splitting node in the entire decision tree. Nodes with higher frequencies usually have greater importance in the decision-making process. Record the number of occurrences and the total number of nodes of each splitting node, calculate its splitting frequency, and store the calculated node splitting frequency data in a structured manner to generate node splitting frequency data. Calculate the degree of splitting discreteness of each node based on the node splitting frequency data. The degree of discreteness measures the distribution range of the node under different splitting conditions. Use statistical methods (such as standard deviation and variance) to calculate the degree of splitting discreteness of each node, generate splitting discreteness data, and provide a basis for weight design. According to the splitting frequency and splitting discreteness of the splitting node, the weight parameters are designed. For example, the greater the discreteness, the smaller the weight parameter. The node with higher weight has greater influence in the decision tree. The weight parameters of each node are calculated using weighted average method or other statistical methods to ensure that the weight reflects the importance and stability of the node. The calculated weight parameters are stored in a structured manner to generate splitting node weight parameters for decision tree model parameter analysis. The decision tree model is analyzed according to the splitting nodes and their corresponding weight parameters. The parameter analysis includes the node splitting conditions, the importance and weight of the splitting nodes, etc. The key parameters of the decision tree, such as node splitting criteria, pruning strategy and weight distribution, are determined to generate decision tree model parameters to ensure that these parameters can accurately guide the construction and optimization of the model. The historical enterprise decision feature data and decision feature probability distribution data are integrated into a comprehensive data set to provide comprehensive data support for decision tree model training. The integrated data set is used to train the model according to the preset decision tree model parameters. During the training process, methods such as cross-validation and grid search are used to optimize model parameters to ensure the accuracy and generalization ability of the model. The effectiveness of the model is verified by evaluating the performance of the model (such as accuracy, precision, recall rate, and F1 score). The trained intelligent decision tree model is saved, and the structure and parameter settings of the model are recorded to ensure that the model can explain the decision-making process, provide visual decision paths and node information, perform performance evaluation, ensure that the model performs consistently on different data sets, and generate the final intelligent decision tree model.

[0184] The beneficial effect of the present application is that the present invention ensures the scientificity and accuracy of the design of the system function module by comprehensively acquiring and finely analyzing the ERP demand data, and the correlation analysis and aggregation processing of the demand data effectively avoids the omission and duplication of requirements, improves the efficiency and accuracy of demand management, and enables the ERP system to better adapt to the actual business needs of the enterprise, and ensures the reasonable configuration and efficient operation of the function module. The microservice architecture mapping and containerization processing of the ERP function module improve the flexibility and scalability of the system. The microservice architecture enables the system to respond to changes in business needs more flexibly, improves the reliability and maintenance efficiency of the system, and the containerization processing ensures the consistency and stability of the system in different environments, simplifies the deployment and operation and maintenance process, and improves the speed and quality of the system going online. The in-depth analysis and processing of historical enterprise source data optimizes data quality and decision-making benefits. The enterprise decision gain based on feature data affects the benefit analysis, provides scientific decision support, improves the accuracy and effectiveness of enterprise decision-making, and helps enterprises make more scientific and efficient decisions in complex business environments. Through the design of intelligent decision tree model and intelligent business practice engine, the intelligence and automation of business processing are realized. The intelligent decision tree model uses historical data and probability distribution information to build an efficient decision model, which improves the accuracy of decision-making. The intelligent business practice engine improves the automation level and adaptability of business processing through intelligent business logic and early warning mechanism, ensuring that enterprises can quickly respond to market changes and business needs. Thus, a well-designed ERP system helps enterprises to carry out overall planning and management.

[0185] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0186] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for launching an ERP system. It is characterized in that The following steps are involved: Step S1: Acquire ERP demand data; perform ERP-related demand analysis based on the ERP demand data to generate ERP-related demand data; Performing associated demand data aggregation processing on ERP associated demand data to generate ERP associated aggregated demand data; Step S1 includes: Step S11: Obtain ERP demand data; Step S12: using natural language technology to perform ERP demand information mining on standard ERP demand data to generate ERP demand mining data; Step S13: performing ERP-related demand analysis on the ERP demand mining data to generate ERP-related demand data; Wherein, step S13 comprises: Step S131: Analyze and process ERP demand entity information according to ERP demand mining data to generate ERP demand entity data; Step S132: performing ERP demand dependency analysis according to the ERP demand entity data to generate ERP demand dependency data; Step S133: using the word frequency-inverse document frequency technology to convert the ERP demand mining data into numerical vectors to generate ERP demand mining vector data; Step S134: performing ERP demand similarity analysis based on the ERP demand mining vector data to generate ERP demand similarity data; Step S135: performing ERP related demand analysis on the ERP demand mining data based on the ERP demand dependency data and the ERP demand similarity data to generate ERP related demand data; Step S14: performing associated demand data aggregation processing on the ERP associated demand data to generate ERP associated aggregated demand data; Step S2: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate ERP-related function modules; ERP system architecture analysis is performed based on the ERP-related function modules to establish the ERP system architecture; Step S3: Acquire historical enterprise source data; remove redundant features from the historical enterprise source data to generate historical enterprise feature data; perform historical enterprise decision feature analysis based on the historical enterprise feature data to generate historical enterprise decision feature data; Step S4: Analyze the enterprise intelligent decision logic parameters according to the historical enterprise decision feature data to generate the enterprise intelligent decision logic parameters; design the intelligent business practice engine based on the enterprise intelligent decision logic parameters to generate the intelligent business practice engine; Wherein, step S4 comprises: Step S41: constructing an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data to generate an intelligent decision tree model; Step S42: Designing enterprise intelligent decision logic parameters based on the intelligent decision tree model to generate enterprise intelligent decision logic parameters; Step S43: designing a preliminary intelligent business practice engine according to the enterprise intelligent decision-making logic parameters; Step S44: Analyze the decision warning indicators of the preliminary intelligent business practice engine through the preset intelligent decision warning threshold, and design the engine decision warning mechanism according to the decision warning indicators; Step S45: Design a supply business smart contract for the preliminary smart business practice engine using blockchain technology, generate a supply business smart contract, and design an engine supply business monitoring mechanism based on the supply business smart contract; Step S46: Perform engine mechanism configuration processing on the preliminary intelligent business practice engine according to the engine decision warning mechanism and the engine supply business monitoring mechanism to generate an intelligent business practice engine; Step S5: Based on the intelligent business practice engine, the intelligent business practice function configuration processing of the ERP system architecture is performed to generate an ERP system, and the ERP system is transmitted to the cloud server to execute the ERP system online operation.

2. The method for launching the ERP system according to claim 1, It is characterized in that Step S2 includes the following steps: Step S21: ERP-related function module analysis is performed according to the ERP-related aggregation demand data to generate an ERP-related function module; Step S22: performing microservice architecture mapping processing of the functional modules according to the ERP-related functional modules to generate microservice architecture mapping data; Step S23: Containerize the microservice architecture mapping data to generate containerized microservice architecture mapping data, and establish an ERP system architecture through the containerized microservice architecture mapping data.

3. The method for launching the ERP system according to claim 1, It is characterized in that Step S3 includes the following steps: Step S31: Obtain historical enterprise source data; Step S32: performing cluster analysis on the historical enterprise source data based on a preset clustering algorithm to generate historical enterprise cluster data; Step S33: Calculate the clustering silhouette coefficient based on the historical enterprise clustering data to generate the clustering silhouette coefficient; Step S34: performing redundant information analysis on the historical enterprise source data according to the clustering silhouette coefficient to generate historical enterprise redundant data; Step S35: Eliminate redundant features of historical enterprise source data according to historical enterprise redundant data to generate historical enterprise feature data; Step S36: Performing an enterprise decision gain impact benefit analysis based on historical enterprise characteristic data to generate enterprise decision gain impact benefit data; Step S37: extracting enterprise decision features of the enterprise decision gain affecting benefits from the historical enterprise feature data according to the enterprise decision gain affecting benefits data, and generating historical enterprise decision feature data.

4. The method for launching the ERP system according to claim 3, It is characterized in that Step S32 includes the following steps: Step S321: Calculate the intra-cluster distance based on the historical enterprise clustering data to generate the intra-cluster distance data; Step S322: Calculate the inter-cluster distance based on the historical enterprise clustering data to generate inter-cluster distance data; Step S323: Calculate the clustering silhouette coefficient based on the intra-cluster distance data and the inter-cluster distance data to generate the clustering silhouette coefficient.

5. The method for launching the ERP system according to claim 3, It is characterized in that Step S36 includes the following steps: Step S361: Evaluate the impact of historical enterprise decisions on benefits based on historical enterprise characteristic data to generate historical enterprise decision impact benefit data; Step S362: Perform enterprise decision gain impact benefit analysis based on historical enterprise decision impact benefit data to generate enterprise decision gain impact benefit data.

6. The method for launching an ERP system according to claim 1, It is characterized in that Step S41 includes the following steps: Step S411: using a preset Bayesian network algorithm to perform decision feature probability distribution analysis on historical enterprise decision feature data to generate decision feature probability distribution data; Step S412: construct an intelligent decision tree model for enterprise decision analysis based on historical enterprise decision feature data and decision feature probability distribution data to generate an intelligent decision tree model.

7. The method for launching an ERP system according to claim 6, It is characterized in that Step S412 includes the following steps: Perform enterprise decision splitting node analysis based on historical enterprise decision feature data to generate enterprise decision splitting nodes; Conduct node splitting frequency analysis on enterprise decision splitting nodes to generate node splitting frequency data; According to the node splitting frequency data, the splitting discrete degree analysis of the enterprise decision splitting node is carried out to generate the splitting discrete degree data; Design the splitting node weight parameters according to the splitting discrete degree data and generate the splitting node weight parameters; Performing decision tree model parameter analysis for enterprise decision analysis based on enterprise decision splitting nodes and splitting node weight parameters to obtain decision tree model parameters; According to the probability distribution data of decision features, the decision tree model parameters are trained for enterprise decision intelligence to generate an intelligent decision tree model.

Images