MIXTURE OF DISTRIBUTION MODELS: APPLICATION TO THE AGE DISTRIBUTION OF PEOPLE WHO COMMITTED SUICIDE

Abstract

Mixture of distribution models are useful tools for analyzing data sets derived from diverse subpopulations, modeled
through a weighted combination of simpler distributions. Various parameters, including mean or shape parameters, may be
estimated using different methods like the EM algorithm or maximum likelihood method. These mixtures aid in modeling
the heterogeneity of multimodal data distributions. The age distribution based on the number of suicides between 2002 and
2020 demonstrates several peaks where the majority of observations are clustered; hence, this modeling approach is
employed for its examination. These data can be represented using a mixture of two-component normal distributions. The
goodness of fit of the model and the estimation of its parameters will support such a result. In conclusion, the aim is to
apply mixture models to the age distribution of people who have committed suicide and identify subgroups that exhibit
specific behaviors or characteristics. This will provide meaningful information to create effective prevention treatments and
aids for different age groups of people at risk.