Graph Communities To Analyze The Occupational Accidents: An Evidence From The Statistics Of Turkey 2013-2014

Tuna G., KURT M.

GAZI UNIVERSITY JOURNAL OF SCIENCE, vol.30, no.4, pp.373-393, 2017 (ESCI) identifier

  • Publication Type: Article / Article
  • Volume: 30 Issue: 4
  • Publication Date: 2017
  • Journal Indexes: Emerging Sources Citation Index (ESCI), Scopus
  • Page Numbers: pp.373-393
  • Gazi University Affiliated: Yes


In this study, we present an efficient graph clustering method called graph communities to analyze occupational accidents in Turkey from 2013 to 2014 and use the Pearson correlation to calculate the similarity measure between the accident data. For this purpose we represent each accident as a vector of a social actor. Then, we obtain sub-communities which can be seen as the cluster of clusters iteratively. Moreover, we present the sub-dominant ultra-metric structures among occupational accidents using the minimum spanning tree method. In each sub-community we analyze the central nodes in minimum spanning trees which are the dominant in the information flow. Furthermore, we use the topological measures to determine which hierarchical structure is best to represent occupational accidents in sub-communities.