Analysis of merchant vessel accidents in Istanbul strait through decision tree method

In this study merchant vessel accidents which occured between 2001 and 2016 in the sectors of Türkeli, Kandilli, Kad?köy, Marmara constitutes Istanbul Strait region under Vessel Traffic Services scope have been examined. Data was obtained from database Ministry Transport Infrastructure Main Search Rescue Coordination Center, after data cleansing process, 535 involve cargo vessels above 500 grosston analyzed. Merchant were taken place specified examined with CHAID (Chi-square Automatic Interaction Detector) Decision Tree method. method is one most common used decision tree algorithms extracting meaningful rules big datasets for classification. Through conducting relationship accident type (collision/contact, grounding other) factors (vessel type, Length overall (LOA), gross tonnage, age, flag, loading condition), time (accident time, season accident) other (sector where occured, pilot on board or not) has Accident occuring sector, board/not, found as important input variables. Based result applied to set, it observed that occurring Kad?köy sector collision / contact 86% probability, Kandilli 48% probability Türkeli both types had 36% probability.