Deep Learning-Based Approach for Human Activity Recognition

Deep learning for the use of human activity detection in smart homes is predicted by the use of this work. Feature extraction and predicting further movement are a challenge, and one-time prediction at a particular time is possible. To accomplish, this RNN with HAR feature extraction is accomplished within the deep learning mechanism. The critical features are extracted at first place, and the rest of the features are extracted only if required. This process identifies the activities at a much faster rate as compared to plain deep learning mechanisms. Classification accuracy of prediction is also better as error rate decays using the HAR feature in deep learning.