Students Activity Recognition through Machine Learning Approaches

Abstract

Movement recognition is currently one of the most wellknown applications of AI calculations. Researchers created many student activity recognition (HAR) systems that turn smartphone readings into various forms of physical activity. It is used in a multitude of industries, including biomedical engineering, game development, and developing more precise metrics for athletic training. Supervised machine learning models can be taught to foresee a person's activities by collecting data from their linked sensors. This project will make use of data from the UCI Machine Learning Repository. This document includes data collected from the phone's multiple sensors, such as the accelerometer, gyrator, and others, and is used to create regulated expectation models using AI techniques like SVM and Arbitrary.

It records data from the phone's sensors, such as the accelerometer and spinner, and uses AI techniques such as assist vector machines and random backwoods to create controlled expectation models. This data can be used to predict six types of development: walking, walking higher up, walking on the bottom floor, sitting, standing, and lying. We will use a confusion matrix to compare the precision of several models.