Algorithm No. of attributes Tree size Time elapsed in second Correctly classified instances Accuracy in % WTPR WFPR WROC J48 28 375 0.48 11291 95.90 96.5 3.5 0.998 26 401 0.45 11262 95.86 96.4 3.7 0.996 24 386 0.38 11281 95.87 96.3 3.7 0.989 22 386 0.37 11281 95.87 96.3 3.7 0.989 20 223 0.34 11230 95.80 96.3 3.7 0.989 16 153 0.23 11111 94.77 94.7 5.3 0.99 Random tree 28 2160 0.08 11274 95.75 96.2 3.9 0.980 26 2450 0.05 11291 95.90 96.3 3.8 0.982 24 2475 0.09 11255 95.59 96.2 3.9 0.981 22 2915 0.11 11280 95.80 96.2 3.8 0.980 20 1368 0.13 11238 95.45 96.0 4.0 0.981 16 909 0.05 11156 94.75 94.8 5.3 0.981 REP tree 28 373 0.27 11291 95.90 96.3 3.7 0.990 26 354 0.42 11286 95.86 96.3 3.7 0.989 24 352 0.22 11284 95.84 96.2 3.8 0.987 22 352 0.22 11284 95.84 96.2 3.8 0.988 20 178 0.36 11253 95.58 96.1 3.9 0.988 16 109 0.13 11152 94.72 94.3 5.7 0.983

Table 10: Performances of J48, Random tree, REP tree classification algorithms on decision tree model building to predict under-five children admission to pediatric ward in NEMM Hospital, 2012.