Research On Collaborative Enterprise Credit Evaluation Method Based On Cloud ERP Ecosystem

Cloud ERP combines emerging technologies such as big data,the Internet and cloud computing to realize the cloud application of ERP,which is a brand-new technology and service model.The cloud ERP platform can support business collaboration between multiple upstream and downstream multi-subject and cross-domain supply chain enterprises,forming a supply chain ecosystem based on the cloud ERP platform.Building a "collaborative" relationship is an essential part of the construction and optimization of the supply chain ecosystem.However,it is difficult to achieve stable and sustainable development of a collaborative relationship that lacks "trust"."Trust" is the necessary premise,and the core factor has become the primary task of collaborative relationship management and optimization in the supply chain ecosystem.Therefore,it is a significant problem that must be solved in the construction of a cloud ERP platform to correctly and effectively evaluate the "trust" relationship of enterprises in the cloud ERP ecosystem,establish a reliable "collaboration" relationship between enterprises,and improve the efficiency of the cloud ERP platform.To this end,this thesis focuses on the national key R＆D program topics "Development and Application Demonstration of Enterprise-level Cloud ERP Platform Supporting Open Ecology"（Project No.: 2019YFB1704104）and "Development and Demonstration Application of Distributed Design,Manufacturing,Operation and Maintenance Platform"（Project No.: The research task of 2021YFB3302104）takes the supply chain collaboration and trust relationship in the Kingdee Cosmic Cloud ERP supply chain ecosystem as the research object and is supported by the enterprise-related data accumulated on the Internet and platforms.As follows:1.Given the problems of incomplete credit evaluation indicators and no applicable data sets,this thesis collects and organizes multi-modal data sets for enterprises in the cloud ERP ecosystem and establishes credit from two aspects of corporate financial indicator data and Internet dynamic comment texts.The evaluation index system has established multi-modal and multi-dimensional data sets related to enterprises in the cloud ERP ecosystem to provide data support for subsequent research.2.For data imbalance,this thesis applies the adversarial generative network to solve the problem of data imbalance.Two different generative adversarial networks are used to generate a few categories of financial indicator data and text review data,respectively,and experiment with mainstream sampling methods.Contrast and demonstrate the effectiveness of generative adversarial generated data.3.Because of the low accuracy of the existing credit evaluation model,when constructing the credit evaluation model,this thesis proposes a Stacking algorithm model based on the attention mechanism,which solves the problem of data feature fusion.Experiments are carried out and compared with existing methods and models to verify the model’s effectiveness proposed in this thesis.4.Given the algorithm’s practical application,this thesis designed and developed a cooperative enterprise credit evaluation tool system.The algorithm model was applied to the ground,and the visual system was convenient for personnel to operate.The system integrates multiple evaluation models for selection.The credit evaluation results can be obtained after inputting enterprise indicators,which significantly improves efficiency and saves indirect enterprise costs.