Summary: | 碩士 === 國立政治大學 === 統計學系 === 87 === A few mathematical models used to fit mortality curve to project future mortality levels are proposed by demographer and statistician in recent years.However, the methods that can be used to appropriately fit the mortality curve in Taiwan have not been brought up yet. Among all other age groups, infants and elderly(60+) mortality change more dramatically than the other age group, as a result the latter has greater impact on the population structure for Taiwan. Hence, our projection focuses on the mortality of elderly population.
This research utilizes the following three fitting methods to project future mortality levels in Taiwan: SOA model; Lee-Carter model; Heligman-Pollard model and the one we proposed namely the "ratio method", along with the regression and time series analyses in order to project the mortality in Taiwan area. Finally, a better projection model is found by comparing the predicted errors between these models.
|