A Novel Ensemble-Based Technique for the Preemptive Diagnosis of Rheumatoid Arthritis Disease in the Eastern Province of Saudi Arabia Using Clinical Data

Table 6

Comparison of the results using backward feature selection.

Features	Classifier	Number of features	Features selected	Test accuracy
Experiment 1
Backward elimination	SVM	16	{Sex, age, MCV, platelet count, MPV, WBC, EOS%, NEU%, LYMPH%, BASO#, SODIUM, EOS#, potassium, chloride, CO2, SGPT}	92.54%
	LR	31	{Sex, age, MCH, MCV, MCHC, RDW, platelet count, MPV, HGB, HCT, WBC, RBC, MONO%, EOS%, NEU%, LYMPH%, BASO%, BASO#, sodium, NEU#, LYMPH#, MONO#, EOS#, potassium, chloride, CO2, anion gap, GGTP, ALP, SGOT, SGPT}	91.04%
	Adaboost	9	{Sex, age, MCHC, RDW, platelet count, MPV, NEU%, potassium, SGPT}	91.04%
	Voting	12	{Sex, age, MCHC, RDW, platelet count, MPV, WBC, MONO%, NEU%, BASO#, potassium, GGTP}	93.28%
Experiment 2
Backward elimination	SVM	22	{Sex, age, MCH, MCV, MCHC, platelet count, MPV, HGB, HCT, RBC, MONO%, EOS%, NEU%, LYMPH%, BASO%, BASO#, sodium, NEU#, potassium, chloride, anion gap, SGPT}	91.79%
	LR	24	{Sex, age, MCH, MCV, MCHC, RDW, platelet count, MPV, HGB, HCT, WBC, RBC, MONO%, EOS%, NEU%, BASO%, BASO#, sodium, MONO#, potassium, CO2, anion gap, GGTP, SGOT}	92.54%
	Adaboost	6	{Sex, age, platelet count, MPV, NEU%, potassium}	92.54%
	Voting	22	{Sex, age, MCH, MCV, MCHC, RDW, platelet count, MPV, HGB, HCT, WBC, RBC, MONO%, EOS%, NEU%, BASO#, sodium, NEU#, LYMPH#, potassium, chloride, ALP}	92.54%