Abstract
We analyzed the social factors of Japanese divorce ratio with using multiple regression methods for prefectural sanitary statistics in2001, and we obtained the regression equations having several explanation variables. In order to investigate characteristics of those ones, we intend to use factor analysis or principal component analysis. On our process, we found that there are some parameters have sometimes a wrong sign like as the relationship between average temperature and snow's days. Those parameters as the work in double harness and academic carrier belong to groups of having wrong signs. If we can practice cluster analysis with using those parameters, most of prefectural and city governments are classified into four groups. However, Tokyo, Okinawa and Hokkaido have nothing to do with groups of other prefectures. We wonder if these three prefectures have the special reason why the other prefectures do so.
