| 摘要 |
<p>PROBLEM TO BE SOLVED: To avoid risks caused by single graphic selection, while overcoming incongruous variance between each axes, human factors based on sight, and arbitrariness. SOLUTION: Mahalanobis distances, using multi-dimensional numerical values, are effectively used in order to overcome incongruous variance between axes, human factors based on sight, and arbitrariness. Sixty graphics are divided into 6 ranks each containing 10 graphics, and comparison results are averaged in each rank to avoid risks caused by dependence upon only a single graphic. Sixty graphics are ranked in the order or proximity to a state image as the examination object, with respect to ordinal average values of the proximity order, based on the Mahalanobis distances from the state image as the examination object, and positions of the graphics are taken as ordinal values to divide the graphics into ranks, and thus they are averaged. In the case a high calculation capability is available, linearly coupled calculation of four kinds of average values is performed, and comprehensive calculations can be effectively used while simplifying calculation result display.</p> |
