Principal Component Analysis of Somatometric Data with Special Reference to the Effect of Selection of Measurement Items

DOI Web Site Web Site

Bibliographic Information

Other Title
  • 生体計測値の主成分分析特に項目の選択の影響について
  • セイタイ ケイソクチ ノ シュセイブン ブンセキ トクニ コウモク ノ センタ

Search this article

Abstract

To examine the effect of adoption or rejection of certain somatometric items upon representativeness and efficiency of a set of somatometric measurements, principal component analysis was repeatedly carried out on the correlation and the variance-covariance matrices derived from 27 somatometric measurements (15 somatic and 12 cephalic) of 17 year-old boys. The purposes of this analysis were: 1) to recognize what sort of variances are contained in the measurements which have been commonly used in somatology; 2) to dis- tinguish the indispensable items from the dependent ones if such a distinction is possible; and 3) to determine which procedure is more favourable to somatological study, calculation from a correlation matrix, or that from a variancecovariance matrix. The conclusions are: 1) it is better to analyse the cephalic items separately from the somatic items, as well as the head items from the face items; 2) main components of the human body build consist of i) development of soft tissues and visscera (robustness), ii) development of the skeletal frame work (especially long bone length), iii) trunk shape, iv) around the chest, and v) around the shoulder, and not all of 15 somatic items are necessary to express these components; 3) there may be some lack and imbalance of informations in the measurements undertaken here; and 4) as far as the present study is concerned, calculation from a correlation matrix has produced better results than calculation from a variance-covariance matrix.

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top