Analysis On Human Activity Recognition Using Machine Learning Algorithm And Personal Activity Correlation

Abstract

Personal Activity Recognition (PAR) has been animportant and interesting problem is solved and also many upgradations are to be added. It is mainly be used for Biometric, Fitness and M healthcare as an assistive technology when ensemble with other technologies likes Machine Learning, Wireless sensor, Internet of Things (IoT) with the help of sensors, smart phones or pictures, PAR can be achieved. We present various up-to-the-minute methods in this study and explain each of them through literature surveys. For each of the methods in which the data is obtained through various means, such as sensors, photographs, accelerometers, gyroscopes, proximity, etc., various datasets are used and the implementation of these machines in different places and the movement of human beings. The results obtained by each and every process and the dataset form are then compared. Machine learning techniques such as decision trees (DT), K-nearest neighbours (KNN), support vector machines (SVM), hidden models are tested for PAR, and deep neural network techniques such as artificial neural networks (ANN), convolution neural networks (CNN) and recurrent neural networks are also examined later.Here, we presented a unified model for finding 95 percent accuracy in the identification of personal behaviour with improvement in the current working application. And we can predict the fitness rate of a specific region among cities, regions and even in a state using this application