Deviation from original tree (lower is better)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.11
Training objective value (lower is better)
0.08
0.09
0.1
0.11
0.12
0.13
0.14
0.15
0.16
0.17
0.18
0.19
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
All trees
Selected tree
Original Tree
≥ 0.7315
< 0.7315
≥ -3.398
< -3.398
≥ 2.259
< 2.259
≥ -0.5387
< -0.5387
≥ 5.167
< 5.167
≥ 1.088
< 1.088
Predict 0
p = 55.54%
1
curtosis
Predict 0
p = 53.26%
2
skewness
Predict 0
p = 58.42%
9
variance
Predict 1
p = 79.29%
3
variance
Predict 0
p = 88.11%
6
variance
Predict 1
p = 83.67%
10
skewness
Predict 0
p = 98.08%
13
Predict 1
p = 96.89%
4
Predict 0
p = 97.30%
5
Predict 1
p = 100.00%
7
Predict 0
p = 100.00%
8
Predict 1
p = 100.00%
11
Predict 0
p = 100.00%
12
×
Tree 22: training error 0.0875, deviation 0.0281
≥ 0.3236
< 0.3236
≥ 2.282
< 2.282
≥ -3.791
< -3.791
≥ -0.2779
< -0.2779
≥ 5.21
< 5.21
Predict 0
p = 55.54%
1
skewness
Predict 1
p = 61.99%
2
curtosis
Predict 0
p = 91.93%
9
variance
Predict 1
p = 94.57%
3
variance
Predict 0
p = 50.60%
6
variance
Predict 1
p = 100.00%
10
Predict 0
p = 100.00%
11
Predict 1
p = 100.00%
4
Predict 0
p = 100.00%
5
Predict 1
p = 91.62%
7
Predict 0
p = 99.35%
8
×