Research On Cotton Processing Quality Improvement Strategy Based On Data Mining

Cotton is an important strategic resource related to the national economy and people's livelihood.Cotton textile enterprises' demand for high-quality lint,competition in the international market,and the adjustment of national policies have made it a top priority to improve the quality of cotton processing.With the development of information technology,"data mining" technology comes into being at the present time.The massive historical data generated in the process of cotton processing contains a lot of instructive information for improving the quality of cotton processing.The data mining technology can be used to mine the potential effective information between cotton processing quality and frequency converter of cotton gin,leather cleaner and seed cleaner,so as to improve cotton processing quality by adjusting frequency converter.The main purpose of this paper is to use data mining technology to put forward corresponding strategies to improve the quality of cotton processing.The main contents include:(1)Preprocessing of cotton historical data.In the process of data collection,various reasons can lead to real data is incomplete or there are outliers.The existence of "dirty data" makes data mining can't directly.So this paper on the history data preprocessing,including regression method is used to fill the missing value,3? criterion to eliminate outliers and Min-Max algorithm standardization history data.(2)Data mining of seed cotton category and lint grade.Different quality improvement strategies need to be provided for different types of seed cotton.In this paper,k-means clustering algorithm is adopted to cluster and classify seed cotton,and two shortcomings of k-means algorithm are improved to avoid the disadvantages of local optimization caused by self-determination of K value and random selection of initial clustering center.Finally,seed cotton is divided into three categories.Due to the diversity and complexity of cotton indicators.There is no unified national standard for the comprehensive quality of cotton.This paper adopts the knowledge of fuzzy grade and membership degree in fuzzy mathematics to develop an empirical model for the intrinsic quality of lint.According to which the comprehensive evaluation index of lint can be calculated,and the cotton can be ranked according to the value.(3)Data mining of lint quality improvement strategy.For each type of seed cotton,association rules were mined for frequency of cotton gin,seed cleaner,leather cleaner and quality index of lint after processing.Due to the basic algorithm of association rules Apriori algorithm is not applicable to numerical data.In this paper,the fuzzy c-means clustering algorithm and the combination of fuzzy association rules.Numerical data can be converted to Boolean data,obtained the cotton processing and cotton gin,Pi Qing machine,and cotton seed cleaning machine inverter frequency synthesis quality.The association rules between the identified the optimal values of the inverter frequency,for each kind of the cotton seed cotton processing quality promotion strategy.(4)Model evaluation and knowledge representation to improve lint quality.In this paper,support vector machine algorithm is used to establish a prediction model for model evaluation of promotion strategy.And grid search method is used to search for kernel function parameter g and penalty factor C of support vector machine,which effectively improves the accuracy of the prediction model.After the model was established,the lint quality under the new strategy was compared with that under the historical strategy.It was proved that 87% of lint quality had been improved.The cotton quality improvement strategy proposed in this paper has certain feasibility.