Smartphone-Based Human Activities Recognition System using Random Forest Algorithm

Abstract

The advancements of smartphone technology, bring together doctors and patients for monitoring clinical activities, remote assistants, and preemptive measures, specifically for those who are under medical control. As smartphones become an integral part of human life, they are employed to monitor health conditions within the human activity recognition (HAR) system. This paper presents a Smartphone-Based Human Activities Recognition System using Random Forest Algorithm(RFA). The RFA algorithm is a classifier that has several decision trees on subclasses of the dataset. The RFA takes the average to enhance the projecting accuracy of the given dataset. Four evaluation parameters such as F1 Score, accuracy, precision, and sensitivity are considered to estimate the proposed system performance. Further, comparing the performance of the proposed RFA with existing classifiers, the accuracy of the proposed system achieves 98.34% as compared with other classifiers.