Abstract:

Sectors in Turkey are clustered according to hazard class is effective in determining the premium rates and hazard ratings for hazard class and the degree of business and workplace. Traditional clustering methods used for his purpose have more limited information because they take into account a single dimension. Clustering these sectors together in terms of similar types of fatal accidents will give detailed results. In order to obtain these results, binary clustering methods developed by traditional clustering methods are discussed. In this study, Bimax algorithm which is one of the biclustering methods is used. In the dataset in the form of matrix covered in the study, the rows indicate the sectors and the columns show the types of the fatal accidents. According to the obtained biclusters, construction and transportation sectors cluster together. The clustering of these sectors is due to the accidents that occurred with motor vehicles and falling from a high place. In this study, sectors and types of accidents were clustered together to form risky sector groups. With this actual practice, measures can be taken to take precautions against fatal accidents by making risk classifications for the sectors.

Keywords: