Comparative Analysis of Generic Outlier Detection Techniques

An anomaly, also known as an outlier, is an event that deviates from the norm and raises suspicion. Such anomalies are found by a procedure called anomaly detection. Every anomaly is a potential threat to the robustness and security of the system, which is why anomaly detection is critical. In this proposed research work, anomaly detection algorithms are implemented to isolate outliers from cured input data. The comparative analysis of the algorithms demonstrates that the Isolation Forest performs better than Gaussian Mixture Model (GMM) and k-Nearest Neighbour (kNN) algorithms.