Package org.apache.mahout.classifier.df.node

Examples of org.apache.mahout.classifier.df.node.Node

217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
    if (alreadySelected) {
      // attribute already selected
      log.warn("attribute {} already selected in a parent node", best.getAttr());
    }

    Node childNode;
    if (data.getDataset().isNumerical(best.getAttr())) {
      boolean[] temp = null;

      Data loSubset = data.subset(Condition.lesser(best.getAttr(), best.getSplit()));
      Data hiSubset = data.subset(Condition.greaterOrEquals(best.getAttr(), best.getSplit()));

      if (loSubset.isEmpty() || hiSubset.isEmpty()) {
        // the selected attribute did not change the data, avoid using it in the child notes
        selected[best.getAttr()] = true;
      } else {
        // the data changed, so we can unselect all previousely selected NUMERICAL attributes
        temp = selected;
        selected = cloneCategoricalAttributes(data.getDataset(), selected);
      }

      // size of the subset is less than the minSpitNum
      if (loSubset.size() < minSplitNum || hiSubset.size() < minSplitNum) {
        // branch is not split
        double label;
        if (data.getDataset().isNumerical(data.getDataset().getLabelId())) {
          label = sum / data.size();
        } else {
          label = data.majorityLabel(rng);
        }
        log.debug("branch is not split Leaf({})", label);
        return new Leaf(label);
      }

      Node loChild = build(rng, loSubset);
      Node hiChild = build(rng, hiSubset);

      // restore the selection state of the attributes
      if (temp != null) {
        selected = temp;
      } else {
View Full Code Here
152
153
154
155
156
157
158
159
160
161
162
   
    log.debug("Building {} trees", nbTrees);
    for (int treeId = 0; treeId < nbTrees; treeId++) {
      log.debug("Building tree number : {}", treeId);
     
      Node tree = bagging.build(rng);
     
      key.set(partition, firstTreeId + treeId);
     
      if (isOutput()) {
        MapredOutput emOut = new MapredOutput(tree);
View Full Code Here
82
83
84
85
86
87
88
89
90
91
92
  void map(IntWritable key, Context context) throws IOException, InterruptedException {
   
    initRandom((InMemInputSplit) context.getInputSplit());
   
    log.debug("Building...");
    Node tree = bagging.build(rng);
   
    if (isOutput()) {
      log.debug("Outputing...");
      MapredOutput mrOut = new MapredOutput(tree);
     
View Full Code Here
70
71
72
73
74
75
76
77
78
79
80
81
        }
      } else if (node instanceof NumericalNode) {
        NumericalNode nnode = (NumericalNode) node;
        int attr = (Integer) fields.get("NumericalNode.attr").get(nnode);
        double split = (Double) fields.get("NumericalNode.split").get(nnode);
        Node loChild = (Node) fields.get("NumericalNode.loChild").get(nnode);
        Node hiChild = (Node) fields.get("NumericalNode.hiChild").get(nnode);
        buff.append('\n');
        for (int j = 0; j < layer; j++) {
          buff.append("|   ");
        }
        buff.append(attrNames == null ? attr : attrNames[attr]).append(" < ")
View Full Code Here
191
192
193
194
195
196
197
198
199
200
201
202
        }
      } else if (node instanceof NumericalNode) {
        NumericalNode nnode = (NumericalNode) node;
        int attr = (Integer) fields.get("NumericalNode.attr").get(nnode);
        double split = (Double) fields.get("NumericalNode.split").get(nnode);
        Node loChild = (Node) fields.get("NumericalNode.loChild").get(nnode);
        Node hiChild = (Node) fields.get("NumericalNode.hiChild").get(nnode);
       
        if (instance.get(attr) < split) {
          buff.append('(').append(attrNames == null ? attr : attrNames[attr])
              .append(" = ").append(doubleToString(instance.get(attr)))
              .append(") < ").append(doubleToString(split));
View Full Code Here
93
94
95
96
97
98
99
100
101
102
103
  @Test
  public void testPredictTrace() throws Exception {
    // build tree
    DecisionTreeBuilder builder = new DecisionTreeBuilder();
    builder.setM(data.getDataset().nbAttributes() - 1);
    Node tree = builder.build(rng, data);
   
    String[] prediction = TreeVisualizer.predictTrace(tree, testData,
        ATTR_NAMES);
    Assert.assertArrayEquals(new String[] {
        "outlook = rainy -> windy = TRUE -> no", "outlook = overcast -> yes",
View Full Code Here
138
139
140
141
142
143
144
145
146
    // build tree
    DecisionTreeBuilder builder = new DecisionTreeBuilder();
    builder.setM(data.getDataset().nbAttributes() - 1);
    builder.setMinSplitNum(0);
    builder.setComplemented(false);
    Node tree = builder.build(rng, lessData);

    assertEquals("\noutlook = sunny\n|   humidity < 85 : yes\n|   humidity >= 85 : no\noutlook = overcast : yes", TreeVisualizer.toString(tree, data.getDataset(), ATTR_NAMES));
  }
View Full Code Here
149
150
151
152
153
154
155
156
157
  public void testEmpty() throws Exception {
    Data emptyData = new Data(data.getDataset());
   
    // build tree
    DecisionTreeBuilder builder = new DecisionTreeBuilder();
    Node tree = builder.build(rng, emptyData);

    assertEquals(" : unknown", TreeVisualizer.toString(tree, data.getDataset(), ATTR_NAMES));
  }
View Full Code Here
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
    if (alreadySelected) {
      // attribute already selected
      log.warn("attribute {} already selected in a parent node", best.getAttr());
    }

    Node childNode;
    if (data.getDataset().isNumerical(best.getAttr())) {
      boolean[] temp = null;

      Data loSubset = data.subset(Condition.lesser(best.getAttr(), best.getSplit()));
      Data hiSubset = data.subset(Condition.greaterOrEquals(best.getAttr(), best.getSplit()));

      if (loSubset.isEmpty() || hiSubset.isEmpty()) {
        // the selected attribute did not change the data, avoid using it in the child notes
        selected[best.getAttr()] = true;
      } else {
        // the data changed, so we can unselect all previousely selected NUMERICAL attributes
        temp = selected;
        selected = cloneCategoricalAttributes(data.getDataset(), selected);
      }

      // size of the subset is less than the minSpitNum
      if (loSubset.size() < minSplitNum || hiSubset.size() < minSplitNum) {
        // branch is not split
        double label;
        if (data.getDataset().isNumerical(data.getDataset().getLabelId())) {
          label = sum / data.size();
        } else {
          label = data.majorityLabel(rng);
        }
        log.debug("branch is not split Leaf({})", label);
        return new Leaf(label);
      }

      Node loChild = build(rng, loSubset);
      Node hiChild = build(rng, hiSubset);

      // restore the selection state of the attributes
      if (temp != null) {
        selected = temp;
      } else {
View Full Code Here
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
    if (alreadySelected) {
      // attribute already selected
      log.warn("attribute {} already selected in a parent node", best.getAttr());
    }

    Node childNode;
    if (data.getDataset().isNumerical(best.getAttr())) {
      boolean[] temp = null;

      Data loSubset = data.subset(Condition.lesser(best.getAttr(), best.getSplit()));
      Data hiSubset = data.subset(Condition.greaterOrEquals(best.getAttr(), best.getSplit()));

      if (loSubset.isEmpty() || hiSubset.isEmpty()) {
        // the selected attribute did not change the data, avoid using it in the child notes
        selected[best.getAttr()] = true;
      } else {
        // the data changed, so we can unselect all previousely selected NUMERICAL attributes
        temp = selected;
        selected = cloneCategoricalAttributes(data.getDataset(), selected);
      }

      Node loChild = build(rng, loSubset);
      Node hiChild = build(rng, hiSubset);

      // restore the selection state of the attributes
      if (temp != null) {
        selected = temp;
      } else {
View Full Code Here
TOP

Related Classes of org.apache.mahout.classifier.df.node.Node

  • org.apache.mahout.classifier.df.builder.DecisionTreeBuilder
  • org.apache.mahout.classifier.df.builder.DefaultTreeBuilder
  • org.apache.mahout.classifier.df.mapreduce.inmem.InMemMapper
  • org.apache.mahout.classifier.df.mapreduce.partial.PartialBuilderTest
  • org.apache.mahout.classifier.df.mapreduce.partial.Step1Mapper
  • org.apache.mahout.classifier.df.tools.TreeVisualizer
  • org.apache.mahout.classifier.df.tools.VisualizerTest
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.