. 2020 Nov 6;20(21):6327. doi: 10.3390/s20216327

Table 3.

Confusion matrix of $e x p e r i m e n t_{1}$ for the user-specific model. Rotation of the bracelet $= N O$ , orientation correction $= N O$ . The $s y n c$ gesture was not used. Classification accuracy $= 94.99 %$ .

	Targets						Predictions Count (Precision%)
	waveIn	waveOut	Fist	Open	Pinch	noGesture	Predictions Count (Precision%)
waveIn	7339	65	73	57	36	168	7738 94.84%
waveOut	86	7416	64	54	32	136	7788 95.22%
fist	18	10	7305	43	19	136	7531 97%
open	79	94	100	7385	113	138	7909 93.37%
pinch	34	41	53	49	7232	150	7559 95.67%
noGesture	94	24	55	62	218	6922	7375 93.86%
Targets Count (Sensitivity%)	7650 95.93%	7650 96.94%	7650 95.49%	7650 96.54%	7650 94.54%	7650 90.48%	45,900 94.99%