Data analysis of failures of signaling and communication equipment in Iranian Railways using data mining techniques

Carrying passenger and freight safely in all aspects to their destinations is of utmost importance for railway administrators. Undergoing safety procedures and developing safety systems requires awareness of what is causing unsafe conditions, which can be done by learning from the past. The department of Signaling & Communication is one of the most important departments in Iranian Railways, which carries out the control of failures in signaling and communication equipment and thus provides the safe condition for train movements. This research has been performed to analyze past failures data of the signaling and communication equipment in Iranian Railway (RAI). For this purpose, some 30,235 failure records from the RAI signaling and communication database were selected from years 2011 and 2012. The Apriori association rule technique, C5.0 decision tree technique, and CRISP-DM methodology are applied. Determining the rules and the relationships among the system attributes and predicting their values, it will be possible to take necessary actions to prevent or mitigate the failures and thus provide the safety in rail transportation.