Efficient Neighborhood Density Based Outlier Detection Inside a Sub Network with High Dimensional Data

Received: 14 November 2018 Accepted: 26 January 2019 Anomaly recognition has been utilized to recognize the exception and remove anomalies from different sorts of information and networks. It has imperative applications in the field of failure recognition, network strength examination, Medical Outlier Detection, Industrial Damage recognition. Detecting few anomalies from a network of information perceptions is a continually testing method. The primary commitment of this work is to build up a technique that can register the neighborhood density based anomalies proficiently in high dimensional information. In this paper, we have demonstrated that the dataset is divided into multiple subsets and checked for exceptions which make the task of outlier detection easy. The exceptions are then consolidated from various subsets. In this way, the neighborhood density based anomalies can be figured effectively. In this paper Density Based Outlier Detection (DBOD) method is proposed which divides the network into sub networks and identifies outliers on them.