Comparative mortality models in Kuwait

This paper compares three different models that have been developed in the literature for modeling and forecasting human mortality rates over an age range. The first model is the quadratic Gompertz model where the quadratic line is fitted for age and sex. The second is the one-dimensional Spline model, which fits linear combinations of the Basis Splines. The last model is the Lee-Carter model which is used to fit each sex to a set of agespecific death rates by fitting the Poisson log-bilinear regression model in generalized linear models (GLM). A time-varying index of mortality is then forecasted using a time series linear forecasting model autoregressive integrated moving average (ARIMA). These forecasts are used to generate projected age-specific mortality rates in Kuwait for ages 55 to 90 for the period 2006-2015, based on the mortality data for years 1993-2005 for males and females separately.