Table 1 Comparison of model performance across WISDM datasets.

Method	Recall	Accuracy	Precision	F1-score
SVM⁶⁰	0.9049	0.9471	0.9258	0.9121
HMM⁶¹	0.9056	0.9436	0.9299	0.9128
Genetic algorithm⁶²	0.9553	0.9724	0.9580	0.9565
GRU	0.9427	0.9676	0.9548	0.9484
GRU-attention⁶³	0.9681	0.9790	0.9740	0.9710
CNN-GRU⁶⁴	0.9377	0.9632	0.9457	0.9414
LSTM	0.9370	0.9600	0.9382	0.9371
Attention-LSTM	0.9398	0.9589	0.9442	0.9419
BiLSTM	0.9593	0.9734	0.9623	0.9607
CNN-LSTM^65,66	0.9557	0.9711	0.9598	0.9576
CNN-BiLSTM⁶⁷	0.9613	0.9740	0.9629	0.9620
CNN-A-BiLSTM⁶⁸	0.9511	0.9656	0.9536	0.9521
CNN-BiGRU⁶⁹	0.9648	0.9755	0.9645	0.9646
TAHAR-student-CNN	0.9880	0.9927	0.9865	0.9872
TAHAR-student-LSTM	0.9327	0.9604	0.9398	0.9342
TAHAR-student-GRU	0.9665	0.9804	0.9745	0.9701
TCN-attention-HAR-teacher	0.9850	0.9903	0.9863	0.9856

Quick links

Search