تحليل شدت مصدوميت ناشي از تصادفها در راههاي دوخطه برون شهري با استفاده از مدل هاي داده كاوي

Analysis of Two-Lane, Two-Way Rural Roads Traffic Injury Severity Based on Data Mining Models

In this study, factors influencing driver and occupant injury severity in two lane , two way roads of Iran are identified. Using statistical models is one of the most common methods that were employed to analyze crash severity. In this study, in addition to examining such models and expressing their weak points, CART method is introduced. Classification and regression trees (CART), which is one of the most common methods of data mining, was employed to analyze the traffic crash data over a three-year period (2006-2008). Expressing models in decision tree format with explicit rules is one of their most useful outputs. In the analysis procedure, the problem of three-class prediction was decomposed into a set of binary prediction models and then eight models were analyzed. This resulted higher overall accuracy of predicting the model, besides the prediction accuracy of the fatality class, which was nearly 0%, and in most of the previous studies, inci eased significantly. Results indicated that improper overtaking and ignoring seat belts are the most important factors affecting the severity of injuries