A heuristic approach for observing outlying points in diabetes data set

Data Mining is the process of analyzing large amount of data and useful for knowledge discovery. Detection of outliers is critically essential in the knowledge based society. Focusing on outlier detection in offline data stream has been increased in the past few years. The proposed a new CLOPD algorithm for identifying mislabelled data (anomaly) during clustering and increase the accuracy of cluster analysis in medical data set It consists of two phases, Partition and Detection. Clustering aims to partitioning the data into groups based on distance metrics. The instance, which does not interfere with respect to the clusters are considered and indentified (Detection) as outliers. The main purpose of the research work is to extracts mislabelled instance (outliers) from data set and its merits are discussed for further exploration. Finally the results were compared and depicted.