The Application Of Data Mining Clustering Algorithm In The Garment Industry

With the continuous development of information technology and emerging of mass data,the companies urgently need efficient,accurate and scientific analysis of data.In recent years,high-dimensional data sets is growing so fast,not a traditional database query methods to extract useful information,so the researchers had to develop new technologies to meet the higher requirements.Currently,data mining technology has gradually entered the practical stage,and will gradually integrate into people's daily lives and bring a variety of services and facilities.Cluster analysis in data mining process,the biggest challenge is not in high-dimensional data set to work,such as microarray gene data,most of the time because of some negative effects and abnormal values lead to inaccurate results,moreover with the increasing of dimension the computational complexity is also increasing rapidly.This article based on current cluster analysis presents a new clustering algorithm which is the Integration of the three algorithms(dimensionality reduction method,principal component analysis and the average distance method algorithms),and explains the algorithm to the actual production,the target market determining and the product improvement in the apparel industry.For the new clustering algorithm,this article did the following related work:First,summarized the current situation of the clustering algorithm in data mining,described the types and applications of the data mining and clustering analysis,furthermore,and compared several major clustering algorithms.Secondly,studied the characteristics of high-dimensional clustering and the main high-dimensional clustering methods,aim at the problems faced for the clustering of high dimensional data,this paper proposed the corresponding solutions.Then,for traditional clustering algorithms cannot meet the status of high-dimensional data,this paper presented a mixture of high-dimensional clustering algorithm based on traditional clustering algorithm,provided the steps of the algorithm and its pseudocode.After that experimental validation of the clustering algorithm by the four data sets,and compared with the traditional clustering methods.Results show that the accuracy of the algorithm increased average of 18.27%,and the execution time reduced by an average 40ms.So,it has better accuracy and efficiency.Finally,this paper provided the feasibility analysis that applied the data mining technology to the apparel industry.By applied this high-dimensional algorithm to the actual production and sales in the garment industry,and clustered customer size data,sales data,and comfort evaluation of data,obtained the Commissioning Specifications Table,Customer Request Form and Clothing Comfort Comparison Table.They are better able guidance of clothing put into operation,target market identification and the product improvement.