Research On Data Mining Of Cotton-spinning Quality

With the rapid development of database technology, and database management system widely used in various applications of the data accumulated more and more data hidden behind the surge in the number of important information, the user wants them to be higher level of analysis, in order to better use of these data. Although the current database system can achieve high efficiency of data entry, query, statistics, and other functions, but can't find the link between data and the information contained, according to the available data can't predict the future trend of development, lack of the data mining of the hidden knowledge , leading to the "explosion of data but the lack of knowledge".The cotton textile industry of China is based on the best of traditional industries in the textile industry .It occupies an important position in the national economy. The development of information technology makes the cotton industry in production management and production process of accumulation of a large number of business data. How to make full use of these data for companies to create efficiency, is looking forward to the current textile enterprises solve important problems, the birth of data mining technology, in order to solve such problems has provided an effective way. How to make full use of these data for companies to create efficiency, is looking forward to be solved important problems of the current textile enterprises .The birth of data mining technology has provided an effective way to solve such problems.This paper takes the qualitative data of Zhejiang Chunjiang cotton spinning enterprise as a foundation, becomes the gauze rank classification question to the cotton spinning enterprise the technical difficulty to carry on the analysis. It introduces the commonly used taxonomic approach like the Decision Tree classification, the Neural Network classification, the Support Vector Fuselage classification, the Genetic algorithm classification and simple Bayesian classify. The Decision Tree sorting algorithm and the simple Bayesian sorting algorithm in becomes in the gauze rank classification question concrete realization was given. In the decision tree sorting algorithm through the computation training regulations in each attribute information gain judgment decision tree each attribute node's order, codes the establishment decision tree by Huffman code, proposed the decision tree cutting out algorithm, refines becomes the rule which in the gauze rank classification uses. It proposed the decision tree cutting out algorithm, refines becomes the rule which in the gauze rank classification uses. The simple Bayesian algorithm is under the independent premise carries on mutually between supposition training regulations each attribute, first calculates each training attribute with to become between the gauze rank the conditional probability, through Bayesian theoretical calculation test data, in each becomes under the gauze rank the conditional probability, the test data will belong to some to become the gauze rank. In this paper, through the contrast between the decision tree and the Bayesian sorting algorithm, has analyzed two kind of sorting algorithm. It proposed the improvement Bayesian sorting algorithm by the foundation of the empirical data.