Feature Signature Discovery for Autism Detection: An Automated Machine Learning Based Feature Ranking Framework

Table 6

Performance of conventional machine learning models with feature selection.

S.No	Dataset	Classifier	FCBF feature signature	Accuracy (%)
1	Toddler	K-NN	A9, A5, A6, A7, A4, A1, A2, A8, and A10	95.6
		RF		95.3
		NB		96.6
		AdaBoost		95.2
2	Child	K-NN	A4, A9, A10, A6, A8, A3, A5, A1, A7, and A2	92.5
		RF		91.4
		NB		96.2
		AdaBoost		92.1

(K-NN : K-nearest neighbor, RF-random forest, NB-naïve Bayes).