Using decision tree-based data mining to establish a sizing system for the manufacture of garments

Abstract

While data mining has been widely used in many fields, little research has been done on its applications to sizing systems for the manufacturing of garments. The goal of this study was to establish systems for using a decision tree technique to determine the pants sizes of army soldiers. We first defined the subject and constructed a large anthropometric database. Second, we prepared and analyzed data, performed factor analyses, and then extracted important sizing variables. Third, we used the decision tree technique to mine data in an effort to identify and classify significant patterns in the body shapes of soldiers. The use of decision tree-based data mining to establish sizing systems is advantageous because it can (1) allow for a wider coverage of body shapes with a fewer number of sizes, (2) generate regular sizing patterns and rules, and (3) provide manufacturers with reference points to facilitate production. The newly developed sizing system can provide garment manufacturers with size specifications, design development, pattern grading, and market analysis. Moreover, when production plans can be made more realistic, inventory costs due to mismatches can be minimized.