ABSTRACT

The determination of probability distributions for modeling mortality rates across ages is a crucial part of mortality analysis. The main objective ofthis study is to obtain a distribution that could yield better fit when used to model the mortality rates pattern of young adults. This study introduced a new continuous distribution using the Kumaraswawamy generator of distributions. The new distribution called Kumaraswamy Ailamujia distribution has three parameters. Statistical properties, namely, quantile function, raw moment,' stochastic ordering and order statistics have been discussed on the basis ofthe distribution. Maximum likelihood estimation of the parameters of the distribution was given the requisite attention. Specifically, the five distributions that were fitted to the data were Gamma, Weibull, Gompertz, Ailamijia and Kummaraswamy Ailamujia distributions. The Road Accident Specific Mortality Rates data were collected from Federal Medical Center, Umuahia, Abia state over the years 1980 -2014. The Road Accident Specific Mortality rates for all the young adult age groups were modeled separately, combined and disaggregated into sex. Goodness offit statistics such as Kolomogorov Smirnov and Grammar Von-Mises Statistic were used to compare the performance of the distributions fitted to the data. The result obtained showed that Kumaraswamy Ailamujia distribution outperformed the other distributions for the combined data across the age groups(10- 29), separate age groups 20-24 and 25-29. For the data of combined age groups 10-29 disaggregated by sex the Kumaraswamy Ailamujia distribution also outperformed other distributions in modeling the male young adult accident cause specific mortality rates. However the Kumaraswamy Ailamujia distribution was outperformed by the Weibul distribution for the female Road Accident Cause Specific Mortality Rates. The Gamma'distribution outperformed the Kumaraswamy Ailazmujia distribution for age group 10-14 and 15-19.