Accurate Anomaly Detection using various Machine Learning methods for IoT devices in Indoor Environment

The Internet of Things i.e. IoT is a collection of specialized devices, or "things" that communicate real-world data through networks. IoT devices are computer devices that can exchange data and connect wirelessly to a network. IPS is a technology that can be used to locate people or items inside buildings using a mobile device such as a smartphone or tablet. Indoor positioning is achieved by using devices already in use such as cellphones, Wi-Fi and Bluetooth antennae, digital cameras, and clocks. RSSI is a metric for assessing the quality of a wireless connection in Wi-Fi devices. But using RSSI values of Wi-Fi may cause inaccuracies in locating any nodes. RSSI is far more prone to interference than a wired network. And these interferences cause the irregularity in Signal strengths. In the case of indoor Wi-Fi positioning, irregular and anomalous RSS data can't be used to pinpoint the location of any unknown node. As a result, this article investigates the RSSI values utilizing machine learning approaches such as supervised, unsupervised, and ensemble learning in an indoor localization scenario employing Wi-Fi devices. The supervised learning techniques utilized in this study include Naive Bayes, Random Forest, and Decision Tree. The SOM (Self-Organizing Map) method was utilized in unsupervised learning. Then, this research work has used stacking method that is used for combining different classifiers implemented for ensemble algorithm. Precision, recall, and f1-score are utilized for experiments. After removing the outliers, the technique was found to be successful, with a high accuracy of 98.2 percent.