Fuzzy Approaches in Forecasting Mortality Rates
Fundamental issues in the study of mortality rate modelling are goodness of fit and the quality of forecasts. These are still open questions despite the fact that dozens of mortality models have been formulated. Capturing all mortality patterns remains elusive for the proposed models. Nevertheless, there are models with better and worse abilities to explain historical mortality rates and to project accurate forecasts. This paper considers two fuzzy approaches for forecasting future mortality rates. First, the fuzzy autoregressive integrated moving average (ARIMA) method allows the making of fuzzy forecasts based on crisp estimates of mortality model parameters. Second, the fuzzy Lee-Carter method models past mortality rates as fuzzy numbers, and then allows the prediction of future fuzzy mortality rates. Numerical findings show that both methods may be useful tools for forecasting.
KeywordsMortality rate Lee-Carter model Fuzzy ARIMA Mortality forecasting
- 1.Box, G.E., Jenkins G.M.: Time Series Analysis: Forecasting and Control, revised edn. Holden-Day (1976)Google Scholar
- 4.Hyndman, R.J., Athanasopoulos, G.: Forecasting: Principles and Practice. OTexts (2014)Google Scholar
- 7.Kosiński, W., Prokopowicz, P., Ślȩzak, D.: Ordered fuzzy numbers. Bull. Pol. Acad. Sci., Ser. Sci. Math 51(3), 327–338 (2003)Google Scholar
- 10.Szymański, A., Rossa, A.: Fuzzy mortality model based on Banach algebra. Int. J. Intell. Technol. Appl. Stat. 7(3), 241–265 (2014)Google Scholar