Table 3 Comparison of model performance across USC-HAD datasets.

Method	Recall	Accuracy	Precision	F1-score
SVM	0.8200	0.8941	0.8641	0.8033
HMM	0.9015	0.9453	0.9199	0.9019
Genetic algorithm	0.9323	0.9556	0.9381	0.9345
GRU	0.7813	0.8461	0.8115	0.7905
GRU-attention	0.9420	0.9581	0.9431	0.9412
CNN-GRU	0.8394	0.8963	0.8485	0.8409
LSTM	0.7663	0.8428	0.7955	0.7709
Attention-LSTM	0.7945	0.8628	0.8252	0.7994
BiLSTM	0.8570	0.8946	0.8644	0.8577
CNN-LSTM	0.8988	0.9347	0.9030	0.8983
CNN-BiLSTM	0.9102	0.9448	0.9062	0.9005
CNN-A-BiLSTM	0.9176	0.9414	0.9110	0.9130
CNN-BiGRU	0.8884	0.9264	0.8860	0.8831
TAHAR-student-CNN	0.8987	0.8976	0.8482	0.8524
TAHAR-student-LSTM	0.8948	0.9317	0.9079	0.8836
TAHAR-student-GRU	0.8976	0.9317	0.8583	0.8701
TCN-attention-HAR-teacher	0.9423	0.9632	0.9488	0.9434

Quick links

Search