Deep Residual Bidir-LSTM for Human Activity Recognition Using Wearable Sensors

<div>Normalized-to-percent matrix confusion on the test dataset using Dee-Res-Bidir-LSTM for the public domain UCI dataset. The columns represent the predicted classes, and the rows represent the actual classes.</div>

Mathematical Problems in Engineering

fig7

Figure 7

Figure 7: Deep Residual Bidir-LSTM for Human Activity Recognition Using Wearable Sensors 