Abstract
Recently, Human Activity Recognition (HAR) is becoming one of the prevalent study fields. HAR is a powerful tool for monitoring a person's dynamism, and it can be accomplished through machine learning (ML) techniques. HAR is a technique of automatically analysing and recognizing human activities depending on information from several wearable devices and smartphone sensors, like location, accelerometer, gyroscope, duration, and other environmental sensors. This study introduces a new Robust Human Activity Recognition using Equilibrium Optimizer with Deep Extreme Learning Machine (RHAR-EODELM) model. The presented RHAR-EODELM technique mainly identifies different classes of human activities. It follows a three-stage process. Initially, the RHAR-EODELM technique employs a min-max normalization process for scaling the activity data. Next, the RHAR-EODELM technique exploits a deep extreme learning machine with a radial basis function (DELM-RBF) model for the prediction process. Finally, the EO approach is enforced to adjust the parameters associated with the DELM-RBF method. A large-scale simulating process highlights the improved HAR results of the RHAR-EODELM method. The experimental values signify that the RHAR-EODELM method reaches improved predictive outcomes over other models.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call