Outlier Detection Using K-Mean and Hybrid Distance Technique on Multi-Dimensional Data Set

Outlier Detection is a major issue in data mining. Outliers are the containments that divert from the other objects. Outlier detection is used to make the data knowledgeable, and easy to understand. There are many type of databases used now days, and many of them contains anomaly objects, detection or removal of these objects is known as outlier detection. In the proposed work outliers are detected by partitioning the dataset with the clustering method that is the K – Mean method using the Mean of Euclidean and Manhattan distance and then find out the outlier with the Hybrid technique that is the mean of the Euclidean and Manhattan Distance. The proposed work is highly efficient in detection of outliers and produces much efficient outliers by using the real bench marked data sets: Iris dataset and Pima Indian Diabetes data set. Index Terms Outliers, Euclidean Distance, Manhattan Distance, Hybrid Technique