Modeling the Cases of Road Traffic Crashes: A Case of Exponential Smoothing Approach
This study was carried out using the monthly (January 2007 to December 2017 with a total of 108 data points) reported road traffic crashes along the roads connecting the main city of Enugu State Nigeria. In this work, we gave a general overview of the three basic aspects of exponential smoothing models proposed with their applications on the cases of road traffic crashes (accident). The relationship between autoregressive integrated moving average models and exponential smoothing models was also given. Using a grid search method for the estimation of smoothing parameters, the best exponential smoothing models for each of the cases was identified based on the Mean square error (MSE), Mean absolute error (MAE) and Mean absolute percentage error (MAPE). The Holt-Winter (Triple) exponential smoothing model expounded by Peter Winter (1960) yielded optimum value in each of the measures for each route since the series exhibit both trend and seasonality.
