Keratoconus Severity Classification Using Features Selection and Machine Learning Algorithms

Table 6

Classification of features selection algorithms based on execution time (hours (h), minutes (min), and seconds (s)) considering 2 and 4 keratoconus classes.

Category	Algorithms	Execution time using 2 classes	Execution time using 4 classes
Category I	Embedded with a filter	1.26 s	1.23 s
	ANOVA	3.59 s	3.59 s
	Mutual information	5.09 s	5.19 s
	Embedded	6.0 s	8.82 s
	Filter with RFE	38.58 s	38.3 s
	Filter with HRFA	1 min 17.58 s	1 min 38.84 s
	RFE	3 min 20.21 s	3 min 19.98 s
Category II	Genetic	9 min 3.07 s	18 min 20.31 s
	Filter with SFS	35 min 54.35 s	45 min 16.61 s
	SFS	3 h 1 min 1.89 s	3 h 31 min 46.47 s
	Filter with SBS	5 h 54 min 3.6 s	7 h 50 min 53.94 s
Category III	SBS	313 h 35 min 16.15 s	N.A