Table 2.

The classification comparation on the UCI and WISDM datasets.

Models	Accuracy (%)	Accuracy (%)
	on UCI	on WISDM
Inherent Features-based CNN4	91.75	91.50
Coherent Features-based CNN5	92.31	92.20
HAR-image CNN7	93.27	91.64
Multi-stage CNN10	94.29	93.19
Attention-based Multi-head CNN9	95.62	94.90
Temporal CNN6	96.31	95.71
Feature Embedding-based LSTM16	96.31	95.55
ConvAE-LSTM21	96.41	95.60
Attention-based LSTM22	95.18	93.74
Hybrid CNN and LSTM17	96.55	95.65
Proposed Method	96.71	95.86