Analysis of Different Classification Techniques for Two-Class Functional Near-Infrared Spectroscopy-Based Brain-Computer Interface

Table 4

Classification accuracy using Naïve Bayes among all subjects.

Feature combination	S1	S2	S3	S4	S5	S6	S7
Mean & slope	95.6	96.9	95.2	96.1	96.4	94.7	96.9
Mean & peak	96.5	98.1	97.1	97.9	96.0	97.9	98.0
Mean & variance	97.5	98.1	96.0	96.6	96.4	95.7	97.1
Slope & peak	94.4	95.0	89.2	95.2	95.0	93.6	96.4
Slope & variance	98.0	98.0	92.0	95.2	94.9	92.0	94.4
Peak & variance	96.9	98.1	92.9	96.9	96.2	92.4	95.4
Peak & skewness	89.6	88.6	82.8	91.8	87.2	88.0	86.0
Mean & skewness	89.3	91.1	89.1	90.6	89.3	87.7	93.2
Slope & skewness	89.5	89.6	83.9	87.8	90.0	86.5	91.3
Kurtosis & skewness	51.1	50.8	51.1	52.7	51.8	55.8	51.9
Variance & skewness	97.7	97.9	90.3	95.6	94.7	90.6	92.1
Peak & kurtosis	89.6	88.2	82.8	92.0	87.1	88.2	86.2
Mean & kurtosis	89.1	91.2	88.8	90.7	89.0	87.6	93.9
Slope & kurtosis	89.6	89.6	83.9	87.7	89.6	85.8	91.2
Variance & kurtosis	97.6	97.9	90.5	95.4	94.5	90.6	92.0