Design And Analysis Of Data Mining Technology For ERP System

With the continuous improvement of science and technology and the increasing popularization of information technology,enterprises have entered the era of information technology.ERP(Enterprise Resources Planning)has been fully and widely used in enterprises.However,in the face of complex business processes and a large number of business data,ERP has not fully applied it,and prone to a phenomenon like "data surge,but lack of knowledge".Data mining technology can solve the above problems better by further analyzing the data,mining out the potential relationship of the data and the hiding information.And it can obtain the key data to support business decisions-making and enhance the competitiveness of enterprises.This thesis analyzes the current status of ERP system and data mining.With the actual project of an enterprise as the background,and on the basis of the existing ERP development of the enterprise,we explored the comprehensive application of ERP and data mining technology related to the material requirement planning.The main research contents are as follows:(1)In order to reduce the interaction between different categories,and mine the connections between the same categories,the KNN(K-Nearest Neighbor)algorithm for classification was improved to optimize the classification determination and improve the accuracy of the algorithm.(2)In order to reduce the inventory accumulation,we need to make sure accurate sales forecast,and make production planning and purchase planning according to sales forecast.The classification results are used as inputs to the BP(Back Propagation)neural network for prediction,reducing the interaction between multiple data and the prediction error.The traditional BP neural network is optimized by L2 regularization and Adam algorithm,further improving the accuracy of model prediction.(3)On the basis of the existing ERP system development status of the enterprise,we according to the actual business needs of the enterprise to design the data mining architecture for ERP.The data mining algorithm of research is integrated with ERP,and the integration architecture of ERP and data mining is designed.(4)The above algorithm is applied to the ERP system in practice,and the implementation effect is given.The research task of this thesis comes from the actual project of the author's enterprise,and the research results have been applied to the ERP system of the enterprise.It can run effectively and obtain the expected results.