直交射影行列を用いた多変量解析 : データ解析の数学的理論

概要

論文の詳細を見る
The purpose of this paper is to represent sorts of the statistical analyses with PQ transforms which are formed by n-column vector 1n. The PQ transforms are the orthogonal projections matrixes which are able to represent the basic statistical values as the orthogonal projections of the variable vector X. In this paper, the PQ transforms were applyed effectively to sorts of the statistical analyses, especially to the canonical correlation analysis, and moreover, some of the statistical analyses were elevated to the multivariate analyse easily. In this sense, I think that applying the PQ transforms to the multivariate analysis were performed in the mathematical expansions successfully. But I neglected applying PQ transforms to the quantification theory type I〜IV, because I would rather have a great interest in the singular value decomposition, the spectral analysis, the inverse problem and so on.
広島文化学園大学の論文
1998-11-01