Data analysis of lightning fatalities occurrence by Non-Homogeneous Poisson Process with applications on public health: The case of Venezuela

Human lightning causalities are undesirable discrete events that occur randomly in space and time and consist on human dangerous situations which are caused by a cloud to ground lightning strike leaving consequences ranging from a temporary or permanent disability condition to a fatality. The statistical analysis of data set of records related to these types of events is a powerful tool for pattern recognition processes on its behavioral that can help to reduce the uncertainty related with its random nature. On this paper is analyzed the time evolution of the occurrence of human lightning fatalities from a mathematical-probabilistic point of view, with the aim of discover time patterns on its behavioral and based on it develop mathematical models with the ability of reduced uncertainty description of the observed time behavioral. The proposed approach is based on the stochastic model know as Non-Homogeneous Poisson Process used in the analysis of stochastic counting and point processes. The statistical data analysis and model fitting are based on a method known as Multi- resolution Analysis. The proposed approach offer important time-type information about the occurrence of lightning fatalities that can help on a better understanding of the incidence of these events on public health.