/
SmoothedAndCurtailedTreeModel.java
144 lines (121 loc) · 5.02 KB
/
SmoothedAndCurtailedTreeModel.java
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
/*
* RapidMiner
*
* Copyright (C) 2001-2008 by Rapid-I and the contributors
*
* Complete list of developers available at our web site:
*
* http://rapid-i.com
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see http://www.gnu.org/licenses/.
*/
package com.rapidminer.operator.learner.meta.thesis;
import java.awt.Graphics;
import java.util.Iterator;
import com.rapidminer.example.Example;
import com.rapidminer.example.ExampleSet;
import com.rapidminer.operator.OperatorException;
import com.rapidminer.operator.learner.SimplePredictionModel;
import com.rapidminer.operator.learner.tree.Edge;
import com.rapidminer.operator.learner.tree.SplitCondition;
import com.rapidminer.operator.learner.tree.Tree;
import com.rapidminer.tools.Renderable;
/**
* The tree model is the model created for decision trees. That can produce
* well calibrated probabilities using "Smoothing" and "Curtailment" as defined
* in the paper "Learning And Decision Making When Costs and Probabilities Are Unkown"
* by Charles Elkan.
*
* @author Michael Green
* @version $Id: SmoothedAndCurtailedTreeModel.java,v 1.8 2008/05/09 19:22:53 greenm9 Exp $
*/
public class SmoothedAndCurtailedTreeModel extends SimplePredictionModel
implements Renderable {
private static final long serialVersionUID = 4368631725370998592L;
private ExtendedTree root;
private final double dBaseRate = 0.05;
private final double dShiftFactor = 200;
private final double dCurtailmentThreshold = 200;
public SmoothedAndCurtailedTreeModel(ExampleSet exampleSet, Tree root) {
super(exampleSet);
this.root = new ExtendedTree(root);
}
public Tree getRoot() {
return this.root;
}
@Override
public double predict(Example example) throws OperatorException {
double dPrediction = predictUsingCurtailment(example, root);
return dPrediction;
}
private double predictUsingCurtailment(Example example, ExtendedTree node) {
double dPrediction = 0;
if (node.isLeaf()) {
dPrediction = predictAsLeaf(example, node);
} else {
boolean bEdgeConditionFound = false;
Iterator<Edge> childIterator = node.childIterator();
while ( (childIterator.hasNext() ) &&
(bEdgeConditionFound == false) ) {
Edge edge = childIterator.next();
SplitCondition condition = edge.getCondition();
if (condition.test(example)) {
bEdgeConditionFound = true;
ExtendedTree childTree = (ExtendedTree) edge.getChild();
if(childTree.getTotalExamples() >= dCurtailmentThreshold)
dPrediction = (predictUsingCurtailment(example, childTree));
else
dPrediction = predictAsLeaf(example, node);
}
}
if(bEdgeConditionFound == false) {
dPrediction = predictAsLeaf(example, node);
}
}
return dPrediction;
}
private double smoothedConfidence(double dPositiveExamples, double dTotalExamples) {
double dSmoothedConfidence = ((dPositiveExamples + (dBaseRate * dShiftFactor))/ (dTotalExamples + dShiftFactor));
return dSmoothedConfidence;
}
private double predictAsLeaf(Example example, ExtendedTree node) {
double dPrediction = 0;
Iterator<String> s = node.getCounterMap().keySet().iterator();
int[] counts = new int[getLabel().getMapping().size()];
int sum = 0;
while (s.hasNext()) {
String className = s.next();
int count = node.getCount(className);
int index = getLabel().getMapping().getIndex(className);
counts[index] = count;
sum += count;
}
for (int i = 0; i < counts.length; i++) {
example.setConfidence(getLabel().getMapping().mapIndex(i), smoothedConfidence(counts[i], sum));
}
dPrediction = getLabel().getMapping().getIndex(node.getLabel());
return dPrediction;
}
public int getRenderHeight(int preferredHeight) {
// TODO Auto-generated method stub
return 0;
}
public int getRenderWidth(int preferredWidth) {
// TODO Auto-generated method stub
return 0;
}
public void render(Graphics graphics, int width, int height) {
// TODO Auto-generated method stub
}
}