Generation and evaluation of distributed cases by clustering of diverse anthropometric data. (2016)

• Record Type: Journal Article
• Authors: Brolin, Erik; Högberg, Dan; Hanson, Lars; Örtengren, Roland
• Is Part Of: International journal of human factors modelling and simulation. Volume 5:Number 3(2016)
• Page Start: 210
• Page End: 229
• Publication Date: 2016
• Subjects: anthropometry -- diversity -- distributed cases -- body size -- body strength -- flexibility -- joint ROM -- range of motion -- capability measurements -- digital human models -- DHM -- modelling -- anthropometric data -- cluster analysis -- hierarchical clustering -- k-means clustering -- Gaussian mixture distribution clustering; Human-computer interaction -- Periodicals; Human-machine systems -- Periodicals; 004.019
• Languages: English
• ISSNs: 1742-5549

Abstract

This paper describes a study where diversity in body size, strength and joint range of motion, together with diversity in other capability measurements, is included in the process of generating data for a group of test cases using cluster analysis. Descriptive statistics and correlation data was acquired for 15 variables for different age groups and both sexes. Based on this data, a population of 10, 000 individuals was synthesised using correlated random numbers. The synthesised data was used in cluster analyses where three different clustering algorithms were applied and evaluated; hierarchical clustering, k-means clustering and Gaussian mixture distribution clustering. Results from the study show that the three clustering algorithms produce groups of test cases with different characteristics, where the hierarchical and k-means algorithm give the most diverse results and where the Gaussian mixture distribution gives results that are in between the first two.