Y. Liu, H. Xu, C. Zhang, X.D. Shi, S. Patnaik

Pages: 113-124

Abstract
Data mining can effectively identify and discover the patterns and inherent laws of accident data. The paper proposes a road traffic accident data mining method based on grey relational clustering. Determine the key influencing factors of road traffic accidents through fault tree analysis method, and achieve accurate quantification of road traffic accident data. Extract the features of road traffic accident data based on EM algorithm. Set the grey relationship analysis factor for road traffic accident data, create a sequence of behavioral feature data, and determine the grey relationship sequence by the operator. Using whitening weight functions to cluster the features of accident data, classify data with consistent features, and achieve road traffic accident data mining. The experimental results show that the designed method has good sensitivity and high grey correlation coefficient in road traffic accident data mining, indicating the feasibility of this method.
Keywords: grey relational clustering; road traffic; accident data mining; fault tree analysis; EM algorithm