Human activity recognition with self-attention

In this paper, a self-attention based neural network architecture to address human activity recognition is proposed. The dataset used was collected using smartphone. contribution of paper multi-layer multi-head for and compared two strong baseline architectures, which are convolutional (CNN) long-short term (LSTM). dropout rate, positional encoding scaling factor also been investigated find the best model. results show that proposed model achieves test accuracy 91.75%, comparable result when both models.