Prediction of accident trend at two-lane federal highways using statistical approach

Road accident is an unfortunate event which is a matter of serious concern to the authority. An accident reduction measure taken in reducing the rate of accidents is by identifying hazardous locations for treatment. The purposes of this study are to determine the blackspot location along Federal Route 1, Federal Route 5 and Federal Route 24 in the Batu Pahat area and to develop an Accident Prediction Model by using multiple regression method as well as to carry out sensitivity analysis on the developed model. This study is limited to the two-lane federal road type and therefore multilane highway is not taken into account. Data for explanatory variables such as volume, speed, and road geometry are obtained using the Metrocount equipment and field measurements. Accident point weightage calculation is carried out in order to determine road accident trends and blackspot ranking using accident data provided by the Royal Malaysian Police. Using these data, an accident prediction model is successfully developed. The results from sensitivity analysis revealed that the increment of accident point weightage can be explained by either the rise in speed, number of major access point, and traffic volume. Meanwhile, an increment of lane width and shoulder width will reduce the weighting point rates.