You are on page 1of 5

Thursday, 10 May 2018 22:26:51 1

The HPSPLIT Procedure

Performance Information

Execution Mode Single-Machine

Number of Threads 2

Data Access Information

Data Engine Role Path

WORK.NEW V9 Input On Client

Model Information

Split Criterion Used Entropy

Pruning Method Cost-Complexity

Subtree Evaluation Criterion Cost-Complexity

Number of Branches 2

Maximum Tree Depth Requested 10

Maximum Tree Depth Achieved 10

Tree Depth 2

Number of Leaves Before Pruning 40

Number of Leaves After Pruning 3

Model Event Level 1

Number of Observations Read 213

Number of Observations Used 213


Thursday, 10 May 2018 22:26:51 2

The HPSPLIT Procedure

Cost-Complexity Analysis for highincome Using Cross Validation


Number of Leaves
1 2 4 7 14 28
0.6

0.5
Average Misclassification Rate

0.4

0.3

0.2

0.1 1–SE
Min Avg Misclass Rate 0.447
N Leaves 2
Parameter 0.0304
0.0

610.33 0.0304 0.0141 0.0056 0.0031 24E-11


Cost-Complexity Parameter
Thursday, 10 May 2018 22:26:51 3

The HPSPLIT Procedure

Classification Tree for highincome

3 4

highincome 1 2
Thursday, 10 May 2018 22:26:51 4

The HPSPLIT Procedure

Subtree Starting at Node=0

Node
N
2

1
2

democracy democracy
1,3 or Miss 2
Node Node
N N
1 2

1 1
2 2

democracy democracy
3 or Miss 1
Node Node
N N
1 2

1 1
2 2

1 highincome=1 2 highincome=2
Thursday, 10 May 2018 22:26:51 5

The HPSPLIT Procedure

Model-Based Confusion Matrix

Predicted

Error
Actual 1 2 Rate

1 55 41 0.4271

2 35 82 0.2991

Model-Based Fit Statistics for Selected Tree

N Mis-
Leaves ASE class Sensitivity Specificity Entropy Gini RSS AUC

3 0.2157 0.3568 0.5729 0.7009 0.8923 0.4315 91.9094 0.6862

ROC Curve for highincome


1.0

0.8

0.6
Sensitivity

0.4

0.2

Training AUC 0.69


0.0

0.0 0.2 0.4 0.6 0.8 1.0


1 - Specificity
Training

Variable Importance

Training

Variable
Variable Label Relative Importance Count

democracy 1.0000 3.6818 2

You might also like