Table 7.

Applied model results for WISDM dataset. MLP: multi-layer perceptron. LR: logistic regression. Stat. Feat.: statistical features. Att. M.: attention mechanism. R. B.: residual block. LSTM: Long short-term memory.

Evaluation	Reference	Segment Length (s)	Feature Extraction	Classifier	Accuracy (%)
10-fold cross validation	Kwapisz et al. [43]	10	Handcrafted	MLP	91.7
	Garcia-Ceja et al. [55]	5	CNN	FC layer	94.2
	Catal et al. [57]	10	Handcrafted	Ensemble of (LR, MLP, j48)	91.62
	Ignatov [58]	10	CNN + Stat. Feat.	FC layer	93.32
	Current model	10	Handcrafted	RF	94
70%/30% split	Gao et al. [56]	10	CNN + Att. M.	FC layer	98.85
	Suwannarat et al. [59]	8	CNN	FC layer	95
	Abdel-Basset et al. [60]	10	CNN + R. B. + LSTM + Att. M.	MLP	98.90
	Zhang et al. [61]	11.2	CNN	FC layers	96.4
	Zhang et al. [62]	10	CNN + Att.	FC layer	96.4
	Current model	10	Handcrafted	RF	98.56