Human Activity Recognition Using Supervised Machine Learning Classifiers

One of the machine learning algorithms’ most popular implementations nowadays is activity recognition. It is used, among other things, in biomedical engineering, the creation of games, and the creation of more precise statistics for athletic training. Supervised machine learning methods can anticipate a person’s tasks using data from connected sensors. In this work, UCI Machine Learning Repository is used for input data. To provide supervised models of prediction, it combines machine learning techniques such decision trees, random forests (RF), GNB, and KNN. The phone’s accelerometer, gyroscope, and other sensors produce it. With this knowledge, it is feasible to predict each movement a person makes, which may be divided into six groups: walking, stair-walking, sitting, standing, and lying. To assess the accuracy of several models, we will utilize a confusion matrix.