抄録
The objective of this study is to improve the reliability of the final diagnosis for a health examination. For this type of study, the inverse matrix of a correlation matrix is calculated, followed by the use of an orthogonal array to select the contributing variables (also called items). In this study however, the Schmidt orthogonalization method was used to select the important variables for reliability improvement and cost reduction. As a result, the Signal-to-Noise ratio increased by 12.6 db after reducing the number of variables from 141 to 79. Naturally, the reliability of diagnosis was improved. Moreover, the Loss Function was used for its economic evaluation.
