A Review on Rail Accidents and Predictions Using Data Mining Techniques

Abstract

The costs of fatalities and injuries from train accidents have a great impact on society. As part of our effort to understand the characteristics of past train accidents, this paper presents an analysis of significant train accidents occurring in around world from 2000 to 2015. Rough set theory and associated rules approaches are applied in analysing the collected data. The results show that although most derived rules are unique, some rules are worth noting. Collision accidents generally lead to more casualties than derailment accidents, and the most frequent cause of accidents is human error. Additionally, most

Train accidents occur during summer. These findings can provide railway leaders with lessons and rules learned from past accidents, thus facilitating the establishment of a safer railway operation environment around world. Accident investigation and analysis are key to reinforcing and improving railway safety. Many railway accidents have been caused by degraded human performance and human error, and the tasks of train drivers and signallers have remained essentially the same. Although new technologies and equipment have gradually reduced railway operation accidents, no investigation has been conducted to investigate whether railway performance shaping factors (R-PSFs), attributed to degraded human performance, have changed or remained constant. The results show that predictive accuracy for accident costs significantly improves through the use of features found by text mining and predictive accuracy further improves through the use of modern ensemble methods. Importantly, this study also shows through case examples how the findings from text mining of the narratives can improve understanding of the contributors to rail accidents in ways not possible through only fixed field analysis of the accident reports.