A New Reachability Based Algorithm for Outlier Detection in Multidimensional Dataset
The quality of data is major role to detect novel results from the large voluminous databases. So the outlier detection is important process in KDD. It is another important area of data mining research. Maximum of the outlier are due to human errors. Many types of outlier detection algorithms are dealt with in the literature. This paper also proposes a new algorithm for detection outliers. It uses new reachability based method for proposed algorithm. The efficiency of the proposed algorithm is proved by using core histogram, letter, segmentation, pima, breast datasets.