-- 作者：c408133217
-- 发布时间：3/9/2012 1:58:00 PM
-- weka functional tree算法的输出结果不明白！！
functional tree和logistic model tree的inner nodes不是一个回归公式吗，为什么weka的输出结果是两个互为负的回归公式。而且最后的leaves node后class1和class2是什么标准分？
下面一个结果的示例，两种模型的weka输出结果类似，所以只放上functional tree的输出结果。
---------------分割线-------------------------------
=== Run information === Scheme:weka.classifiers.trees.FT -I 15 -F 0 -M 15 -W 0.0
Relation: training-weka.filters.unsupervised.attribute.Remove-R1-weka.filters.unsupervised.attribute.Remove-R3-10,34-1171,1228-1302-weka.filters.unsupervised.attribute.Remove-R22,36-38,40,45,49-50,59,62,66-77,79
Instances:204
Attributes:59
MW
AMW
nAT
nSK
nBT
nBO
nBM
SCBO
nCIC
nCIR
RBN
RBF
nDB
nTB
nAB
nH
nC
nN
nO
nP
nS
nCL
nBR
nX
nCp
nCs
nCt
nCq
nCrH2
nCrHR
nCrR2
nCaH
nCaR
nCconjR
nNCO
nCOOHPh
nCOOR
nCOORPh
nCONHR
nCONR2
nOCON
nCOXPh
nCO
nCONN
nNH2
nNH2Ph
nNHR
nNHRPh
nNR2
nNR2Ph
nCN
nNO2Ph
nOH
nOHPh
nOHp
nPhX
nHDon
nHAcc
judgement
Test mode:10-fold cross-validation === Classifier model (full training set) === FT tree
------------------ N0#1 <= 0.538508
| N0#2 <= 0.188679: FT_1:15/45 (76)
| N0#2 > 0.188679: FT_2:15/45 (32)
N0#1 > 0.538508: FT_3:15/30 (96) Number of Leaves : 3 Size of the Tree : 5
FT_N0#1:
Class 0 :
-1.45 +
[MW] * 0 +
[nCIR] * 0.1 +
[nDB] * -0.14 +
[nN] * 0.5 +
[nS] * 0.36 +
[nCp] * 0.17 +
[nCrH2] * 0.07 +
[nCrHR] * 0.21 +
[nCrR2] * 0.43 +
[nCaR] * 0.07 +
[nCOORPh] * -0.58 +
[nNH2Ph] * 0.47 +
[nNR2] * -0.28 +
[nPhX] * 0.28 Class 1 :
1.45 +
[MW] * 0 +
[nCIR] * -0.1 +
[nDB] * 0.14 +
[nN] * -0.5 +
[nS] * -0.36 +
[nCp] * -0.17 +
[nCrH2] * -0.07 +
[nCrHR] * -0.21 +
[nCrR2] * -0.43 +
[nCaR] * -0.07 +
[nCOORPh] * 0.58 +
[nNH2Ph] * -0.47 +
[nNR2] * 0.28 +
[nPhX] * -0.28 FT_N0#2:
Class 0 :
-2.22 +
[MW] * 0.01 +
[nCIR] * 0.1 +
[nDB] * -0.14 +
[nN] * 0.5 +
[nS] * 0.36 +
[nBR] * 1.52 +
[nCp] * 0.17 +
[nCq] * 1.56 +
[nCrH2] * 0.18 +
[nCrHR] * 0.56 +
[nCrR2] * 1.95 +
[nCaR] * 0.07 +
[nCOOR] * -0.93 +
[nCOORPh] * -0.58 +
[nCONHR] * -0.54 +
[nCONR2] * 1.52 +
[nCOXPh] * -0.74 +
[nNH2] * -0.77 +
[nNH2Ph] * 0.47 +
[nNR2] * -0.28 +
[nNR2Ph] * 1.31 +
[nPhX] * 0.8 +
[nHAcc] * -0.13 Class 1 :
2.22 +
[MW] * -0.01 +
[nCIR] * -0.1 +
[nDB] * 0.14 +
[nN] * -0.5 +
[nS] * -0.36 +
[nBR] * -1.52 +
[nCp] * -0.17 +
[nCq] * -1.56 +
[nCrH2] * -0.18 +
[nCrHR] * -0.56 +
[nCrR2] * -1.95 +
[nCaR] * -0.07 +
[nCOOR] * 0.93 +
[nCOORPh] * 0.58 +
[nCONHR] * 0.54 +
[nCONR2] * -1.52 +
[nCOXPh] * 0.74 +
[nNH2] * 0.77 +
[nNH2Ph] * -0.47 +
[nNR2] * 0.28 +
[nNR2Ph] * -1.31 +
[nPhX] * -0.8 +
[nHAcc] * 0.13 FT_1:
Class 0 :
-1.92 +
[MW] * 0.01 +
[nBT] * -0.06 +
[nCIR] * 0.1 +
[RBF] * 4.33 +
[nDB] * -0.45 +
[nH] * -0.27 +
[nN] * 0.5 +
[nS] * 2.88 +
[nBR] * 1.52 +
[nCp] * 1.08 +
[nCq] * 1.56 +
[nCrH2] * 0.18 +
[nCrHR] * 0.56 +
[nCrR2] * 1.95 +
[nCaR] * 0.07 +
[nCOOHPh] * 1.16 +
[nCOOR] * -0.93 +
[nCOORPh] * -0.58 +
[nCONHR] * -0.54 +
[nCONR2] * 1.52 +
[nCOXPh] * -0.74 +
[nNH2] * -0.77 +
[nNH2Ph] * 0.47 +
[nNR2] * -0.28 +
[nNR2Ph] * 1.31 +
[nNO2Ph] * 1.8 +
[nPhX] * 0.8 +
[nHDon] * -1.19 +
[nHAcc] * -0.13 Class 1 :
1.92 +
[MW] * -0.01 +
[nBT] * 0.06 +
[nCIR] * -0.1 +
[RBF] * -4.33 +
[nDB] * 0.45 +
[nH] * 0.27 +
[nN] * -0.5 +
[nS] * -2.88 +
[nBR] * -1.52 +
[nCp] * -1.08 +
[nCq] * -1.56 +
[nCrH2] * -0.18 +
[nCrHR] * -0.56 +
[nCrR2] * -1.95 +
[nCaR] * -0.07 +
[nCOOHPh] * -1.16 +
[nCOOR] * 0.93 +
[nCOORPh] * 0.58 +
[nCONHR] * 0.54 +
[nCONR2] * -1.52 +
[nCOXPh] * 0.74 +
[nNH2] * 0.77 +
[nNH2Ph] * -0.47 +
[nNR2] * 0.28 +
[nNR2Ph] * -1.31 +
[nNO2Ph] * -1.8 +
[nPhX] * -0.8 +
[nHDon] * 1.19 +
[nHAcc] * 0.13 FT_2:
Class 0 :
-0.83 +
[MW] * 0.01 +
[nBT] * -0.03 +
[nCIR] * -0.03 +
[nDB] * -0.14 +
[nN] * -0.65 +
[nS] * 0.36 +
[nBR] * 1.52 +
[nCp] * 0.17 +
[nCq] * 2.29 +
[nCrH2] * 0.18 +
[nCrHR] * 0.56 +
[nCrR2] * 1.95 +
[nCaH] * 0.05 +
[nCaR] * 0.07 +
[nCconjR] * 0.7 +
[nCOOHPh] * -0.64 +
[nCOOR] * -2.8 +
[nCOORPh] * -0.58 +
[nCONHR] * -0.54 +
[nCONR2] * 1.52 +
[nCOXPh] * -2.31 +
[nCO] * 0.36 +
[nNH2] * -0.77 +
[nNH2Ph] * 0.47 +
[nNHR] * 1.57 +
[nNR2] * -0.28 +
[nNR2Ph] * 2 +
[nOH] * -1.16 +
[nPhX] * 0.8 +
[nHAcc] * -0.13 Class 1 :
0.83 +
[MW] * -0.01 +
[nBT] * 0.03 +
[nCIR] * 0.03 +
[nDB] * 0.14 +
[nN] * 0.65 +
[nS] * -0.36 +
[nBR] * -1.52 +
[nCp] * -0.17 +
[nCq] * -2.29 +
[nCrH2] * -0.18 +
[nCrHR] * -0.56 +
[nCrR2] * -1.95 +
[nCaH] * -0.05 +
[nCaR] * -0.07 +
[nCconjR] * -0.7 +
[nCOOHPh] * 0.64 +
[nCOOR] * 2.8 +
[nCOORPh] * 0.58 +
[nCONHR] * 0.54 +
[nCONR2] * -1.52 +
[nCOXPh] * 2.31 +
[nCO] * -0.36 +
[nNH2] * 0.77 +
[nNH2Ph] * -0.47 +
[nNHR] * -1.57 +
[nNR2] * 0.28 +
[nNR2Ph] * -2 +
[nOH] * 1.16 +
[nPhX] * -0.8 +
[nHAcc] * 0.13 FT_3:
Class 0 :
0.02 +
[MW] * 0 +
[nBT] * -0.02 +
[nCIR] * 0.1 +
[RBF] * -1.76 +
[nDB] * 0.05 +
[nN] * 0.5 +
[nP] * -0.89 +
[nS] * 1.28 +
[nCL] * -0.16 +
[nCp] * 0.17 +
[nCs] * 0.03 +
[nCrH2] * 0.07 +
[nCrHR] * 0.21 +
[nCrR2] * 0.43 +
[nCaR] * 0.07 +
[nCOORPh] * -0.58 +
[nOCON] * -0.89 +
[nNH2Ph] * 0.06 +
[nNR2] * -0.78 +
[nNO2Ph] * 0.34 +
[nOHPh] * -0.31 +
[nPhX] * 0.28 Class 1 :
-0.02 +
[MW] * 0 +
[nBT] * 0.02 +
[nCIR] * -0.1 +
[RBF] * 1.76 +
[nDB] * -0.05 +
[nN] * -0.5 +
[nP] * 0.89 +
[nS] * -1.28 +
[nCL] * 0.16 +
[nCp] * -0.17 +
[nCs] * -0.03 +
[nCrH2] * -0.07 +
[nCrHR] * -0.21 +
[nCrR2] * -0.43 +
[nCaR] * -0.07 +
[nCOORPh] * 0.58 +
[nOCON] * 0.89 +
[nNH2Ph] * -0.06 +
[nNR2] * 0.78 +
[nNO2Ph] * -0.34 +
[nOHPh] * 0.31 +
[nPhX] * -0.28
Time taken to build model: 0.44seconds === Stratified cross-validation ===
=== Summary === Correctly Classified Instances 157 76.9608 %
Incorrectly Classified Instances 47 23.0392 %
Kappa statistic 0.5389
Mean absolute error 0.2359
Root mean squared error 0.4548
Relative absolute error 47.3282 %
Root relative squared error 91.1099 %
Total Number of Instances 204 === Detailed Accuracy By Class === TP Rate FP Rate Precision Recall F-Measure ROC Area Class
0.759 0.219 0.796 0.759 0.777 0.771 N
0.781 0.241 0.743 0.781 0.761 0.771 R
Weighted Avg. 0.77 0.229 0.771 0.77 0.77 0.771 === Confusion Matrix === a b <-- classified as
82 26 | a = N
21 75 | b = R
|