Table 8.

Classification methods applied to the PAMAP2 dataset.

Method	Description	Accuracy	F1-Score
CNNs [43]	CCNs are used for feature extraction from acceleration time  series.	91%	91.16%
LSTMs [43]	The performance of LSTMs for real-time HAR is analyzed and compared with some other DL/ML models.	85.86%	85.34%
LSTMs [44]	Temporal and sensor attentions are added to LSTMs to improve their performance for HAR.	-	89.96%
BiLSTMs [43]	BiLSTMs are applied to the real-time HAR domain.	89.52%	89.4%
CNN-LSTMs [45]	The HAR performance of CNNs and that of CNN-LSTMs are compared.	88.68%	88.98%
SVMs [43]	The application of SVMs to real-time HAR is investigated and their performance is compared to some other ML/DL models.	84.07%	83.76%
SVMs [46]	class-based decision fusion is used for effective combination of sensor data.	-	82.32%
KNNs [47]	A feature extraction technique is proposed for accelerometer data recorded by sensors in smart devices.	-	91.1%