主成分, 反復数量化計算, 交互最小二乗法, 選択簡便法
A variable selection method using criteria in Tanaka and Mori's modified principal component analysis selects a reasonable subset of quantitative variables that provides principal components which are computed using only a selected subset of variables but which represent all of the variables, including those not selected, as much as possible.
In this paper modified principal component analysis is extended so that it can deal with qualitative data with unordered or ordered categories. Namely both quantification of qualitative data and modified principal component analysis are performed at the same time. To do this the iteration technique based on the alternating least squares method by Young et al.'s PRINCIPALS is used for quantification at every step of finding the best subset.
Example data is analyzed to demonstrate the performance and usefulness of the proposed method, in which four cost-saving selection procedures are compared for real data analysis.
