In today’s world we have a positive trend of better infrastructure with huge amount of roads expansion and variety of vehicles moving around the globe. As the number of vehicles is increasing, more number of commuters involved there is a probability of increased number of accidents which can happen. In this paper Canopy K Means algorithm is implemented on the data provided by government of India and then states are classified into Low, Medium and High Accident Zones. Unlike normal k means algorithm the proposed method select the optimized centers so that better classification results are obtained. During the clustering process instead of working on all the attributes of accident data sets top level attributes are found out based on standard deviation.

Keywords : K Means, Eigen Value, Clustering.