Application of multivariate analysis techniques in the study of accident risk associated with the characteristics of vehicles in the city of Quito

Abstract

 In this paper we analyze if the physical characteristics associated to the vehicles of the city of Quito influence the occurrence of accidents. For this, a sample of 62,474 insured vehicles was taken and the multivariate statistical techniques of Multiple Correspondence Analysis and Logistic Regression were used, to study the relationships between the physical factors of the vehicles and the occurrence of a casualty. With the help of the statistical software IBM SPSS 22, it was determined that the brand, type and color of the vehicle are characteristics that affect the probability of occurrence of a loss.