Predicting the Number of People for Road Traffic Accident on Highways by Hour of Day

摘要

In the present, the volume of vehicles has become crucial factor for the number of road accidents Then, this paper is predicting the number of people for traffic accident by hour of day regarding deaths and injuries on highways in Thailand. The predicting number of people are separated into two types: (i) predict the total number of people in both of deaths and injuries and (ii) predict the number of male and female in deaths and injuries separately. Time series predicting methods: Multiplicative Decomposition, Multiplicative Holt-Winters, and Seasonal Autoregressive Moving Integrated Moving Average (SARIMA) model are applied to predict the number of people deaths and injuries. The Predicting error is measured by using the mean absolute percentage error (MAPE). For this study, separately predicting number of male and female resulted in less predicting errors than predicting the total number of people in female type of deaths and injuries. However, the predicting total number of people method is appropriated to predicting the number of male for both deaths and injuries. With this predict, the transport and traffic policy department can determine the strategy for road traffic accidents for any period.