A Collaborative Outlier Detection Method for High Dimensional Data sets

In the popularization of data mining techniques there is increasing demand among the researchers and academicians to detect the ramified database of different kinds that deals with areas such as healthcare sector, insurance, banking etc. Many researchers proposed different methods which were distance based approaches, density based approaches, depth based approaches etc. but these approaches were not able to deal properly with high dimensional datasets. This paper proposes a collaborative outlier detection approach for high dimensional dataset with four phases say Normalization phase, Reduction phase, Clustering phase, and lastly Feature bagging phase. The experiment was conducted on lung Cancer dataset obtained from UCI repository and the outcomes are recorded, moreover the outcome shows that detection accuracy improves considerably and simultaneously there is decrease in the execution time by five percent.