The Research And Application On Data Mining Technology In Lean Production

With the faster development of information technology and net technology and the globalization of economy, the enterprises, especially the manufactures turn to face the continuously increasing press of competition. To survive and develop under this market environment, the enterprises must innovate and transform continuously at the aspect of production technology and management fashion. Combining with the computer application technology, enhancing the production decision level, increasing the flexibility of production, shortening of the delivery period and decreasing the cost are very important for improving of the enterpriser's competitive power.At present, in the research field of the mode of Chinese textile production ,the intellectualized production decision-making is applied, all kinds of excellent intelligent arithmetic have brought about some effects. However, most of them linger about in aspects of modeling and application, as to the optimizing selection of arithmetic parameters, it haven't been applied in engineering, for this reason it is a unsolved and critical problem to find out the high-efficient and intelligent optimization algorithm.Setting textile industry as background, quality prediction as research object, optimization algorithm as main contents, artificial intelligence and information technology as the tool, the paper pay more attention to the quality prediction system that can effectively help the enterprise solve the parameters optimization and choice. Firstly, the paper analyze the data mining and its application in the system, and makes the demand analysis about the Small-Scale Enterprise. Secondly, it introduce data mining theory, and according to feature of the textile processing, analyses three algorithm models: ANN,GA-SVM,Decision tree. Then, based on quality data of the three type (cotton spinning,wool spinning,chemical fiber) ,to compare the traditional model with the genetic algorithm optimization model, the stability of the GA-SVM model is verified. At the last, ground on COM, it makes a solution to GA-SVM model, run the system and gets the effect analysis.In this paper, the research results are GA-SVM model and technical method, which based on data-mining theory depend on genetic-algorithm optimization which support vector machine preferences for quality-forecasting system .the method takes the advantages of GA and SVM, it work out the complex preferences by ingenious algorithm design, and it gives full play to MATLAB and C# in program development and improves the efficiency for program development and operation and has some reference values.