Interpolation Method for Synthesized Household Compositions Keeping Household Continuity

In this paper, we propose an interpolation method for synthesized household compositions between census years by keeping household continuity. In order to develop a digital twin for human societies, the information on household compositions are essential. Since the census is conducted every five years in Japan, the current synthetic population for Japan does not include data for years when the census is not conducted. For these gap years between census years, there is a need to generate interpolated data to reflect the demographic trends during those periods. Miyatake & Murata interpolated data at five-year intervals based on population transition data for years. However, in that study, the continuity of households was not enough. Therefore, in order to solve issues in their study, we will resynthesize the synthetic population over a five-year period, taking into account remaining households and moving population. Remaining households are determined from census statistics and households are replaced to minimize errors with the statistics for each year. The interpolated data generated by this study reflect more accurately actual population dynamics and will be useful for future population projections and policy making.