Two-Stage Bagging Pruning for Reducing the Ensemble Size and Improving the Classification Performance

Table 10

The highest classification accuracy values on 28 datasets when using four types of base classifiers DT, GNB, KNN, and LR.

Dataset	DT	GNB	KNN	LR
Aba	55.09	52.55	53.72	55.37
Adult	85.35	81.93	81.99	84.04
Aus	87.54	80.58	69.86	86.52
Bcw	96.42	93.85	94.71	94.13
Bld	73.62	62.03	66.09	63.77
Cmc	52.89	50.51	51.26	51.12
Col	87.50	66.30	82.88	80.98
Cre	86.81	82.32	79.71	83.77
Der	96.72	95.36	97.81	98.36
Ger	77.30	74.00	68.90	77.40
Gla	74.30	41.12	75.70	61.21
Hea	82.22	85.19	80.37	81.48
Hep	83.87	70.97	83.87	81.94
Ion	94.59	90.31	84.90	86.61
Kr-vs-kp	99.62	64.99	96.40	91.18
Mam	78.46	79.40	79.50	75.65
Pid	76.72	76.59	74.51	76.98
Spe	82.77	74.91	76.40	79.03
Tel	87.97	72.87	83.08	79.03
Veh	76.24	46.81	66.90	63.36
Vot	95.86	94.71	95.40	94.25
Vow	90.81	70.10	95.66	52.32
Yea	60.78	44.34	56.87	55.39
Spambase	94.63	82.48	91.22	89.11
Tictacto	97.18	70.67	84.45	70.77
Wdbc	97.36	94.38	93.50	95.25
Wpbc	76.26	76.26	76.26	77.27
Spect	81.65	67.04	82.40	84.27