KNN Based Outlier Detection Algorithm in Large Dataset

An outlier is the object which is very different from the rest of the dataset on some measure. Finding such exception has received much attention in the data mining field. In this paper, we propose a KNN based outlier detection algorithm which is consisted of two phases. Firstly, it partitions the dataset into several clusters and then in each cluster, it calculates the Kth nearest neighborhood for object to find outliers. In addition, the pruning scheme is used in our algorithm. It can effectively avoid frequent passing the entire dataset and unnecessary computations. Experimental results on both synthetic and real life datasets show that our algorithm is efficient for outlier detection in large dataset.