INVESTIGATION INTO MACHINE LEARNING TECHNIQUES FOR NOVELTY DETECTION IN WIRELESS SENSOR DATA

Abstract

Since many years ago, anomaly detection has been utilized to locate and separate abnormal components from data. Anomalies have been found using a variety of ways. Machine Learning (ML), which is one of the increasingly important approaches, is crucial in this domain. In order to determine which model works best or how to increase accuracy by making changes to the cur-rent model so that it can adapt to different datasets, this research compares the performance of various machine learning and deep learning models for outlier detection on the IBRL (Intel Berkeley Research Lab) dataset and find which model suits the best or how to improve the accuracy by making changes in the existing model so that the model could adapt to various datasets also.