Forecasting of Road Accident in Kerala: A Case Study

Traffic accidents are the foremost reason for death and injuries around the world, fatalities are still on the ascent in many creating nations including India. Data Analysis of road accidents makes a strong impact for taking preventive measure to overcome the mishap. In this manuscript, we have addressed the prediction problem of road accidents using time series analysis across all districts of Kerala. Time series analysis is useful in discovering the trends in road accidents which enables the prediction of future patterns. In the present MS, we used the time series road accidents data in Kerala, India for the period January 1999 - December 2016 to understand the patterns in the data and to develop appropriate model to predict about future patterns which may enable authorities to take preventive steps. We subsetted the data till 2013 December as training data for the model selection and rest of the data is used for model valuation. Two models discussed here are the “Holt-Winters (HW) exponential smoothing” and “Seasonal ARIMA (SARIMA)”. Both the models will provide the forecast values within the confidence interval of the test data.