A logistic regression analysis on the influence of accident factors on the fatalities of road accidents in metro Manila

Abstract

Using binary logistic regression, the data collected from Metropolitan Manila Development Authority (MMDA) were examined to determine if accident factors contribute to accident fatality. A total of 29,409 accidents were studied after tallying all the accidents with the complete information on all the considered factors. The classification (fatal or nonfatal) is the dependent variable while the other factors such as age, gender, junction-type, weather, location, time and vehicle-type are the independent variables. The model formulated in this study shows the relationship of the dependent variable and the significant factors in terms of odds ratio concept. The findings show that all the independent variables except for age are statistically significant in the accident fatality of road accidents in Metro Manila from 2006 to 2016.

Keywords: Binary logistic regression, Odds ratio, Probabilities, Regression coefficients, Regression model