OUTLIER DETECTION USING INNER AND OUTER RADIUS BASED METHOD

Outlier detection is a fundamental issue in data mining, specifically it has been used to detect and remove anomalous objects from data mining. The proposed approach to detect outlier includes two distances which are inner radius and outer radius. Inner radius is calculated from the global centroid distance to the nearest cluster distance minus the radius of that cluster. Similarly we calculate outer radius which is the maximum distance between global centroid and any one of the cluster plus that cluster radius. For clustering FCM algorithm is used which partition the dataset into given number of clusters. The clustering is done only on useful data points. This will act as a model of my project on the basis of these clusters we will point out outlier. These two radius we will point out the outlier points. While pointing any point to be an outlier we will also check, are there any groups of points which form another cluster, for that case we have to check that condition separately. Those points which are outside outer radius are outlier points and those points which are less then inner radius are outlier points.