Combination Prediction Model of Traffic Accident Based on Rough Set

Abstract

Vehicles accidents have become the first public nuisance in the world. A dramatic rise of traffic accidents results from sharp increase of vehicles with the rapid development of economy. Accident forecasting is designed to help decision-making and planning before casualty and loss occur. Calculating weight coefficient is a key for combination forecast. The result of the forecast will be straightly influenced if the selection of the weighting coefficient is illogicality. A new method of combination forecasting applied in traffic accident is showed in this paper. It is based on the rough sets theory, and the weighting coefficient of all the forecast methods is distributed, so that the calculation of the weighting coefficient will be more impersonal and simpler, and the result of the forecast will be more exactly. In this paper, two samples were used to check the accuracy of this method. The Percent of errors were approximately about 0.5%and 2.7%. Compared with another method for combination forecasting- artificial neutral network, the Percent of errors were 1.1% and 3.05%. Respectively.