Fast clustering of male lower body based on GA-BP neural network

Abstract

PurposeThe purpose of this paper is to improve the prediction accuracy of the body shape prediction model and provide some reference value for the design of underwear.Design/methodology/approachThe body size data of 250 male youths is measured to analyze the body shape of the lower body. And there is a total of 56 measurement items, which are clustered by GA-BP-K-means, K-means, optimal segmentation method for ordered samples, wavelet coefficient analysis, regression analysis and Naive Bayes Algorithm. Finally, a test male sample of an unknown body shape was clustered to verify the superiority of the GA-BP-K-means.FindingsThis paper presented the key factors for body shape clustering, and experimental results have shown that the GA-BP neural network model is higher in speed and precision than other algorithm prediction models.Originality/valueIt was clarified which is the key to body shape clustering. At the same time, the GA-BP-K-means algorithm can promote the popularization and application of the prediction model in body shape clustering.