| 1 | /* |
|---|
| 2 | * This program is free software; you can redistribute it and/or modify |
|---|
| 3 | * it under the terms of the GNU General Public License as published by |
|---|
| 4 | * the Free Software Foundation; either version 2 of the License, or |
|---|
| 5 | * (at your option) any later version. |
|---|
| 6 | * |
|---|
| 7 | * This program is distributed in the hope that it will be useful, |
|---|
| 8 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|---|
| 9 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|---|
| 10 | * GNU General Public License for more details. |
|---|
| 11 | * |
|---|
| 12 | * You should have received a copy of the GNU General Public License |
|---|
| 13 | * along with this program; if not, write to the Free Software |
|---|
| 14 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
|---|
| 15 | */ |
|---|
| 16 | |
|---|
| 17 | /* |
|---|
| 18 | * TwoWayNumericSplit.java |
|---|
| 19 | * Copyright (C) 2001 University of Waikato, Hamilton, New Zealand |
|---|
| 20 | * |
|---|
| 21 | */ |
|---|
| 22 | |
|---|
| 23 | package weka.classifiers.trees.adtree; |
|---|
| 24 | |
|---|
| 25 | import weka.core.Instance; |
|---|
| 26 | import weka.core.Instances; |
|---|
| 27 | import weka.core.RevisionUtils; |
|---|
| 28 | import weka.core.Utils; |
|---|
| 29 | |
|---|
| 30 | import java.util.Enumeration; |
|---|
| 31 | |
|---|
| 32 | /** |
|---|
| 33 | * Class representing a two-way split on a numeric attribute, of the form: |
|---|
| 34 | * either 'is < some_value' or 'is >= some_value'. |
|---|
| 35 | * |
|---|
| 36 | * @author Richard Kirkby (rkirkby@cs.waikato.ac.nz) |
|---|
| 37 | * @version $Revision: 1.6 $ |
|---|
| 38 | */ |
|---|
| 39 | public class TwoWayNumericSplit |
|---|
| 40 | extends Splitter { |
|---|
| 41 | |
|---|
| 42 | /** for serialization */ |
|---|
| 43 | private static final long serialVersionUID = 449769177903158283L; |
|---|
| 44 | |
|---|
| 45 | /** The index of the attribute the split depends on */ |
|---|
| 46 | private int attIndex; |
|---|
| 47 | |
|---|
| 48 | /** The attribute value that is compared against */ |
|---|
| 49 | private double splitPoint; |
|---|
| 50 | |
|---|
| 51 | /** The children of this split */ |
|---|
| 52 | private PredictionNode[] children; |
|---|
| 53 | |
|---|
| 54 | /** |
|---|
| 55 | * Creates a new two-way numeric splitter. |
|---|
| 56 | * |
|---|
| 57 | * @param _attIndex the index of the attribute this split depeneds on |
|---|
| 58 | * @param _splitPoint the attribute value that the splitter splits on |
|---|
| 59 | */ |
|---|
| 60 | public TwoWayNumericSplit(int _attIndex, double _splitPoint) { |
|---|
| 61 | |
|---|
| 62 | attIndex = _attIndex; |
|---|
| 63 | splitPoint = _splitPoint; |
|---|
| 64 | children = new PredictionNode[2]; |
|---|
| 65 | } |
|---|
| 66 | |
|---|
| 67 | /** |
|---|
| 68 | * Gets the number of branches of the split. |
|---|
| 69 | * |
|---|
| 70 | * @return the number of branches (always = 2) |
|---|
| 71 | */ |
|---|
| 72 | public int getNumOfBranches() { |
|---|
| 73 | |
|---|
| 74 | return 2; |
|---|
| 75 | } |
|---|
| 76 | |
|---|
| 77 | /** |
|---|
| 78 | * Gets the index of the branch that an instance applies to. Returns -1 if no branches |
|---|
| 79 | * apply. |
|---|
| 80 | * |
|---|
| 81 | * @param inst the instance |
|---|
| 82 | * @return the branch index |
|---|
| 83 | */ |
|---|
| 84 | public int branchInstanceGoesDown(Instance inst) { |
|---|
| 85 | |
|---|
| 86 | if (inst.isMissing(attIndex)) return -1; |
|---|
| 87 | else if (inst.value(attIndex) < splitPoint) return 0; |
|---|
| 88 | else return 1; |
|---|
| 89 | } |
|---|
| 90 | |
|---|
| 91 | /** |
|---|
| 92 | * Gets the subset of instances that apply to a particluar branch of the split. If the |
|---|
| 93 | * branch index is -1, the subset will consist of those instances that don't apply to |
|---|
| 94 | * any branch. |
|---|
| 95 | * |
|---|
| 96 | * @param branch the index of the branch |
|---|
| 97 | * @param instances the instances from which to find the subset |
|---|
| 98 | * @return the set of instances that apply |
|---|
| 99 | */ |
|---|
| 100 | public ReferenceInstances instancesDownBranch(int branch, Instances instances) { |
|---|
| 101 | |
|---|
| 102 | ReferenceInstances filteredInstances = new ReferenceInstances(instances, 1); |
|---|
| 103 | if (branch == -1) { |
|---|
| 104 | for (Enumeration e = instances.enumerateInstances(); e.hasMoreElements(); ) { |
|---|
| 105 | Instance inst = (Instance) e.nextElement(); |
|---|
| 106 | if (inst.isMissing(attIndex)) filteredInstances.addReference(inst); |
|---|
| 107 | } |
|---|
| 108 | } else if (branch == 0) { |
|---|
| 109 | for (Enumeration e = instances.enumerateInstances(); e.hasMoreElements(); ) { |
|---|
| 110 | Instance inst = (Instance) e.nextElement(); |
|---|
| 111 | if (!inst.isMissing(attIndex) && inst.value(attIndex) < splitPoint) |
|---|
| 112 | filteredInstances.addReference(inst); |
|---|
| 113 | } |
|---|
| 114 | } else { |
|---|
| 115 | for (Enumeration e = instances.enumerateInstances(); e.hasMoreElements(); ) { |
|---|
| 116 | Instance inst = (Instance) e.nextElement(); |
|---|
| 117 | if (!inst.isMissing(attIndex) && inst.value(attIndex) >= splitPoint) |
|---|
| 118 | filteredInstances.addReference(inst); |
|---|
| 119 | } |
|---|
| 120 | } |
|---|
| 121 | return filteredInstances; |
|---|
| 122 | } |
|---|
| 123 | |
|---|
| 124 | /** |
|---|
| 125 | * Gets the string describing the attributes the split depends on. |
|---|
| 126 | * i.e. the left hand side of the description of the split. |
|---|
| 127 | * |
|---|
| 128 | * @param dataset the dataset that the split is based on |
|---|
| 129 | * @return a string describing the attributes |
|---|
| 130 | */ |
|---|
| 131 | public String attributeString(Instances dataset) { |
|---|
| 132 | |
|---|
| 133 | return dataset.attribute(attIndex).name(); |
|---|
| 134 | } |
|---|
| 135 | |
|---|
| 136 | /** |
|---|
| 137 | * Gets the string describing the comparision the split depends on for a particular |
|---|
| 138 | * branch. i.e. the right hand side of the description of the split. |
|---|
| 139 | * |
|---|
| 140 | * @param branchNum the branch of the split |
|---|
| 141 | * @param dataset the dataset that the split is based on |
|---|
| 142 | * @return a string describing the comparison |
|---|
| 143 | */ |
|---|
| 144 | public String comparisonString(int branchNum, Instances dataset) { |
|---|
| 145 | |
|---|
| 146 | return ((branchNum == 0 ? "< " : ">= ") + Utils.doubleToString(splitPoint, 3)); |
|---|
| 147 | } |
|---|
| 148 | |
|---|
| 149 | /** |
|---|
| 150 | * Tests whether two splitters are equivalent. |
|---|
| 151 | * |
|---|
| 152 | * @param compare the splitter to compare with |
|---|
| 153 | * @return whether or not they match |
|---|
| 154 | */ |
|---|
| 155 | public boolean equalTo(Splitter compare) { |
|---|
| 156 | |
|---|
| 157 | if (compare instanceof TwoWayNumericSplit) { // test object type |
|---|
| 158 | TwoWayNumericSplit compareSame = (TwoWayNumericSplit) compare; |
|---|
| 159 | return (attIndex == compareSame.attIndex && |
|---|
| 160 | splitPoint == compareSame.splitPoint); |
|---|
| 161 | } else return false; |
|---|
| 162 | } |
|---|
| 163 | |
|---|
| 164 | /** |
|---|
| 165 | * Sets the child for a branch of the split. |
|---|
| 166 | * |
|---|
| 167 | * @param branchNum the branch to set the child for |
|---|
| 168 | * @param childPredictor the new child |
|---|
| 169 | */ |
|---|
| 170 | public void setChildForBranch(int branchNum, PredictionNode childPredictor) { |
|---|
| 171 | |
|---|
| 172 | children[branchNum] = childPredictor; |
|---|
| 173 | } |
|---|
| 174 | |
|---|
| 175 | /** |
|---|
| 176 | * Gets the child for a branch of the split. |
|---|
| 177 | * |
|---|
| 178 | * @param branchNum the branch to get the child for |
|---|
| 179 | * @return the child |
|---|
| 180 | */ |
|---|
| 181 | public PredictionNode getChildForBranch(int branchNum) { |
|---|
| 182 | |
|---|
| 183 | return children[branchNum]; |
|---|
| 184 | } |
|---|
| 185 | |
|---|
| 186 | /** |
|---|
| 187 | * Clones this node. Performs a deep copy, recursing through the tree. |
|---|
| 188 | * |
|---|
| 189 | * @return a clone |
|---|
| 190 | */ |
|---|
| 191 | public Object clone() { |
|---|
| 192 | |
|---|
| 193 | TwoWayNumericSplit clone = new TwoWayNumericSplit(attIndex, splitPoint); |
|---|
| 194 | clone.orderAdded = orderAdded; |
|---|
| 195 | if (children[0] != null) |
|---|
| 196 | clone.setChildForBranch(0, (PredictionNode) children[0].clone()); |
|---|
| 197 | if (children[1] != null) |
|---|
| 198 | clone.setChildForBranch(1, (PredictionNode) children[1].clone()); |
|---|
| 199 | return clone; |
|---|
| 200 | } |
|---|
| 201 | |
|---|
| 202 | /** |
|---|
| 203 | * Returns the revision string. |
|---|
| 204 | * |
|---|
| 205 | * @return the revision |
|---|
| 206 | */ |
|---|
| 207 | public String getRevision() { |
|---|
| 208 | return RevisionUtils.extract("$Revision: 1.6 $"); |
|---|
| 209 | } |
|---|
| 210 | } |
|---|