An anomaly is something that deviates from normal or from what we expect. Anomalies exist in every field Credit card data, Medical diagnosis, and Law enforcement and so on. What we need is a method for detecting anomalies in data so we can easily find a way to solve this. Here we are using a two stage approach for anomaly detection. In this two stage hybrid approach we use K-Means and Depth Based Outlier Detection Model and K-Means, Deviation Based Outlier Detection Model. Where we state that the hybrid approach stated here is much more efficient than the existing hybrid approach using k-means and Density based. The k-Means performs the clustering of the data and each of the algorithms, namely deviation based and Depth based is used for anomaly detection where we do the detailed study of the clusters to detect anomaly. By this two stage approach Anomaly detection is done in an efficient manner. © Research India Publications.
cited By 0
Aa Donna, Shilpa, Ma, and K.G.b Gaina, “Hybrid approach for Anomaly Detection”, International Journal of Applied Engineering Research, vol. 10, pp. 2170-2174, 2015.