Use of Machine Learning Algorithms for the Prediction of Abnormal Behavior in Humans

Abstract

As an example, smoking at a petrol station is obviously not a good idea and might put people's lives in jeopardy. The goal of this research is to identify the most effective Machine Learning method for dealing with such prediction issues. Smoking, calling, and typical behaviors make up the datasets gathered for behavior detection. Linear Support Vector Machine (LSVM), Kernel Support Vector Machine (KSVM), Decision Tree Classifier (DT), Random Forest Classifier (RF), K-nearest Neighbors (KNN), and K-Means Clustering are some of the well-known algorithms studied in these experiments. Mean Squared Error (MSE) and Confusion Matrix are also used to evaluate the efficiency of each method. Lastly, PCA (Principal Component Analysis) shows how the optimal technique worked. With an accuracy of 82%, Random Forest Classifier (RF) outperforms all other methods and can predict when individuals will act in an aberrant way. Component-wise keywords: ML, dimensionality reduction, abnormal behavior prediction