Research Article

A Novel Multiway Splits Decision Tree for Multiple Types of Data

Table 8

The accuracy of five classifiers on numerical data.

DatasetAccuracy % (rank)
J48CARTSLOC1HHCART(A)MSDT

Blood78.090.50(1)70.51  0.70(5)77.34  0.56(3)76.50  0.45(4)77.53  0.68(2)
Glass66.73  2.84(3)66.59  2.06(4)68.43  2.80(2)64.02 3.14(1)
Image96.55  0.26(3)96.32  0.39(4)96.78  0.33(2)97.220.25(1)96.01  0.40(5)
Iris95.13  0.73(3)95.07  0.61(4)94.67  0.62(5)95.33  0.63(2)96.000.00(1)
Letter88.09  0.23(5)88.19  0.17(4)89.72  0.20(2)88.23  0.22(3)91.770.16(1)
MF-mor72.010.62(1)65.11  0.43(5)70.15  0.48(3)69.83  0.63(4)71.13  0.43(2)
Wav176.50  0.32(4)75.38  0.52(5)80.07  0.33(2)79.83  0.33(3)82.300.32(1)
Wav275.29  0.61(4)74.27  0.43(5)79.68  0.56(2)79.01  0.45(3)80.650.42(1)
Wine93.54  0.88(3)90.00  1.46(5)92.54  0.98(4)94.47  0.86(2)95.111.23(1)
Yeast55.77  0.93(3)51.14  1.01(5)55.901.06(1)54.28  1.11(4)55.84  0.93(2)
Average79.77(3)77.26(4.6)80.53(2.6)79.87(3.1)81.68(1.7)