Classification And Prediction Of Spare Parts Based On Fuzzy Clustering And Correlation Analysis

China has the world’s largest semiconductor market, under the stimulus of smart terminal, consumer electronics, cloud computing, big data, networking and other emerging industries, the semiconductor market demand will continue to grow. However, due to the level of development of the semiconductor industry and advanced countries there is still a big gap, the majority of semiconductor products rely on imports, more than oil and commodities to become the first major import commodities.In recent years, with the push of market demand and national policy support, the semiconductor industry development rapidly, which semiconductor packaging and testing industry due to the IC Design of technical barriers and wafer fabrication of financial barriers to become the development focus, accounting for semiconductor industry total output value of more than 40%. In the semiconductor packaging and testing industry, most of the investment funds are used to purchase equipment, the normal operation of the equipment is the guarantee of continuous production activities of the production enterprise. In the process of equipment maintenance and repair, need to use the spare parts to replace damaged spare parts, so the importance of equipment spare parts is self-evident. Scientific management of spare parts can ensure the timely supply of spare parts, reduce inventory costs, reduce the cost of capital and avoid unnecessary waste, thereby enhancing the competitiveness of enterprises in the complex business environment.On the basis of summarizing the existing research results, first of all, this paper introduces the concept of data mining and the application of data mining briefly, data mining process model and related algorithms. Subsequently, it analysis the spare parts management of packaging and testing companies in spare parts classification and demand forecast problems existing in two aspects:classification and prediction methods are too simple, considering single factor, the subjectivity is stronger, and unable to meet the management decision-making analysis based on the example of a semiconductor packaging and testing companies. Then combined with the data mining, the data mining analysis model of spare parts management is put forward. For spare parts classification problem, the fuzzy clustering method based on multi dimension index is adopted. The demand forecasting is based on the traditional forecasting model. Then, taking the spare parts data of a semiconductor package as an example, considering the economy, supply and demand of spare parts, the fuzzy clustering of spare parts is divided into four categories by using FCM algorithm. Based on the classification of spare parts, according to the historical demand time series data, the Apriori algorithm is applied to the analysis of the association rules of spare parts demand, and combined with GM (1,1) and other forecast models to forecast the demand of spare parts. Finally, based on MATLAB and ASP.net, the data mining analysis module of spare parts management system is realized. The system can provide timely decision-making information for management decision makers.