An Efficient and Effective Model to Handle Missing Data in Classification

Table 1

Specifications of real-world datasets.

Dataset name	Sample size	Variable number	Discrete variable number	Missing proportion	Imbalance
Breast Cancer Wisconsin [47]	699	10	0	2.29	65.5
Chronic kidney disease	400	24	13	60.5	62.5
Congressional voting records	435	16	16	46.67	61.4
Credit approval	690	15	9	5.36	55.5
Cylinder bands	540	39	19	48.7	57.8
Heart disease—ungarian	294	13	7	99.66	63.9
Hepatitis	155	19	13	48.39	79.4
Horse colic	368	23	15	98.1	63
Mammographic mass [48]	961	5	2	13.63	53.7
Ozone level detection	2536	73	0	27.13	97.1