Bed Position Classification by a Neural Network and Bayesian Network Using Noninvasive Sensors for Fall Prevention

Table 7

Comparison of sleep position classification algorithms.

Ref	# of positions	Accuracy (%)	Algorithm	Type of sensors	# of sensors
[13]	8	97.1	kNN	Pressure sensors	2,048
[14]	3	98.4	GMM+kNN	Pressure sensors	1,728
[15]	5	97.7	PCA+SVM	Pressure sensors	360
[16]	5	98.1	HoG+DNN	Pressure sensors	2,048
[17]	4	99.7	SVM	Pressure sensors	512
[18]	5	97.7	kNN	Force sensing array	2048
[19]	6	83.5	Raw data+SVM	FSR sensors	56
[20]	9	94.1	Joint feature extraction and normalization+SVM+PCA	FSR sensors/video	60
[21]	3	100	Kurtosis+skewness	FSR sensors	16
[22]	5	98.4	SVM (linear)+SVM (RBF)+LDA	CC-electrodes	12
Ours	3	97.8	NN+Bayesian network	Pressure and piezoelectric sensors	4