2021 Volume 56 Issue 3 Pages 1282-1288
In this study, we applied the Bayesian APC analysis, using the population data of each prefecture and age group of the national census from 1980 to 2015. It was observed that the age effect was maximum in six prefectures; particularly, individuals in their 20s were continuously flowing in and out. We classified the date based on the age effect—younger type, elderly type, employment type—which is related to residence for employment and schooling; the period effect—increasing type, decreasing type, convex type, concave type—which is related to the characteristics of the regional economic situation; and the cohort effect—old generation type, new generation type—which is related to the characteristics of people born in the same age group. It was found that the characteristics of population changes in 47 prefectures during the survey period can be classified into 17 types.
