Factor Analysis for Finding Invariant Neural Descriptors of Human Emotions

Table 1

Test Accuracy of 4 classifiers (NN, SVM, RF, and kNN) with feature selection methods: ESN: Echo State Network; IFE: Interval Feature Extraction; ICA: Independent Component Analysis; GPC: Graded Probabilistic Clustering; OFS: Optimal Feature Selection; Factor Analysis (FA).

	-	ESN	IFE	ICA	GPC	OFS	FA
NN	75,0	85,9	78,1	81,8	83,5	80,3	90.4
SVM	58,3	86,7	77,1	72,9	64,2	73,6	90.4
RF	50,0	79,7	71,3	69,4	62,8	65,2	84.6
kNN	56,7	83,3	72,2	72,6	76,8	79,5	88.5