Package org.fnlp.ml.types

Examples of org.fnlp.ml.types.InstanceSet

116
117
118
119
120
121
122
123
124
125
126
    long starttime = System.currentTimeMillis();
    // 将样本通过Pipe抽取特征
    Pipe pipe = createProcessor();

    // 测试集
    testSet = new InstanceSet(pipe);

    testSet.loadThruStagePipes(new SequenceReader(testfile,hasLabel,"utf8"));
    System.out.println("测试样本个数：\t" + testSet.size()); // 样本个数

    long featuretime = System.currentTimeMillis();
View Full Code Here
200
201
202
203
204
205
206
207
208
209
210
    long beginTime = System.currentTimeMillis();

    Pipe pipe = createProcessor();

   
    trainSet = new InstanceSet(pipe, factory);
     labels = factory.DefaultLabelAlphabet();
    features = factory.DefaultFeatureAlphabet();
    features.setStopIncrement(false);
    labels.setStopIncrement(false);
View Full Code Here
240
241
242
243
244
245
246
247
248
249
      if (!hasTarget) {// 如果test data没有标注
        pipe = featurePipe;
      }

      // 测试集
      testSet = new InstanceSet(pipe);
      testSet.loadThruStagePipes(new SequenceReader(testfile, hasTarget, "utf8"));
      System.out.println("Test Number: " + testSet.size()); // 样本个数
    }
  }
View Full Code Here
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
    //建立pipe组合
    SeriesPipes pp = new SeriesPipes(new Pipe[]{ngrampp,targetpp,indexpp});
   
    SeriesPipes  pp2 = new SeriesPipes(new Pipe[]{prePipe, ngrampp,targetpp,indexpp});
   
    InstanceSet instset = new InstanceSet(pp2,af);



    //读入数据，并进行数据处理
    instset.loadThruStagePipes(reader);

    /**
     * 建立分类器
     */   
    OnlineTrainer trainer = new OnlineTrainer(af,100);
View Full Code Here
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
      ARClassifier tc = new ARClassifier();
      tc.train();
      Linear cl =Linear.loadFrom(modelFile);
      int i = 0;int j = 0;double ij = 0.0;int kk = 0;int jj = 0;int nn = 0;int n = 0;
      InstanceSet test = new InstanceSet(cl.getPipe(),cl.getAlphabetFactory());
      SimpleFileReader sfr = new SimpleFileReader("../tmp/ar-train.txt",true);
     
      ArrayList<Instance> list1 = new ArrayList<Instance>();
      while (sfr.hasNext())
      {
        list1.add(sfr.next());
      }
      List<String>[] str1 = new List[list1.size()];
      String[] str2 = new String[list1.size()];
      Iterator it = list1.iterator();
      while(it.hasNext()){
        Instance in = (Instance) it.next();
        str1[i] = (List<String>) in.getData();
        str2[i] = (String) in.getTarget();
        i++;
      }
      for(int k = 0;k<str2.length;k++)
      {
        if(str2[k].equals("1"))
          kk++;
      }
      String ss =null;
        test.loadThruPipes(new ListReader(str1));
       
        for(int ii=0;ii<str1.length;ii++){
          ss = cl.getStringLabel(test.getInstance(ii));
          if(ss.equals("1"))
            j++;
       
          if(ss.equals("1")&&ss.equals(str2[ii]))
            jj++;
View Full Code Here
138
139
140
141
142
143
144
145
146
147
148
149
      //构造转换器组
      SeriesPipes pipe = new SeriesPipes(new Pipe[]{lpipe,fpipe});



      InstanceSet instset = new InstanceSet(pipe,factory);
      instset.loadThruStagePipes(new SimpleFileReader(trainFile," ",true,Type.LabelData));
      Generator gen = new SFGenerator();
      ZeroOneLoss l = new ZeroOneLoss();
      Inferencer ms = new LinearMax(gen, factory.getLabelSize());
      Update update = new LinearMaxPAUpdate(l);
      OnlineTrainer trainer = new OnlineTrainer(ms, update,l, factory.getFeatureSize(), 50,0.005f);
View Full Code Here
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
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
    Pipe targetpp = new Target2Label(af.DefaultLabelAlphabet()); 
    //建立pipe组合
    SeriesPipes pp = new SeriesPipes(new Pipe[]{removepp,segpp,s2spp,targetpp,sparsepp});

    System.out.print("\nReading data......\n");
    InstanceSet instset = new InstanceSet(pp,af); 
    Reader reader = new MyDocumentReader(trainDataPath,"gbk");
    instset.loadThruStagePipes(reader);
    System.out.print("..Reading data complete\n");
   
    //将数据集分为训练是和测试集
    System.out.print("Sspliting....");
    float percent = 0.8f;
    InstanceSet[] splitsets = instset.split(percent);
   
    InstanceSet trainset = splitsets[0];
    InstanceSet testset = splitsets[1]; 
    System.out.print("..Spliting complete!\n");

    System.out.print("Training...\n");
    af.setStopIncrement(true);
    BayesTrainer trainer=new BayesTrainer();
    BayesClassifier classifier= (BayesClassifier) trainer.train(trainset);
    System.out.print("..Training complete!\n");
    System.out.print("Saving model...\n");
    classifier.saveTo(bayesModelFile); 
    classifier = null;
    System.out.print("..Saving model complete!\n");
    /**
     * 测试
     */
    System.out.print("Loading model...\n");
    BayesClassifier bayes;
    bayes =BayesClassifier.loadFrom(bayesModelFile);
    System.out.print("..Loading model complete!\n");
   
    System.out.println("Testing Bayes...");
    int flag=0;
    float[] percents_cs=new float[]{1.0f,0.9f,0.8f,0.7f,0.5f,0.3f,0.2f,0.1f};
    int[] counts_cs=new int[10];
    for(int test=0;test<percents_cs.length;test++){
      System.out.println("Testing Bayes"+percents_cs[test]+"...");
      if(test!=0)
        bayes.fS_CS(percents_cs[test]);
      int count=0;
      for(int i=0;i<testset.size();i++){
        Instance data = testset.getInstance(i);
        Integer gold = (Integer) data.getTarget();
        Predict<String> pres=bayes.classify(data, Type.STRING, 3);
        String pred_label=pres.getLabel();
        String gold_label = bayes.getLabel(gold);
       
        if(pred_label.equals(gold_label)){
          count++;
        }
        else{
          flag=i;
//          System.err.println(gold_label+"->"+pred_label+" : "+testset.getInstance(i).getTempData());
//          for(int j=0;j<3;j++)
//            System.out.println(pres.getLabel(j)+":"+pres.getScore(j));
        }
      }
      counts_cs[test]=count;
      System.out.println("Bayes Precision("+percents_cs[test]+"):"
      +((float)count/testset.size())+"("+count+"/"+testset.size()+")");
    }
    bayes.noFeatureSelection();
    float[] percents_csmax=new float[]{1.0f,0.9f,0.8f,0.7f,0.5f,0.3f,0.2f,0.1f};
    int[] counts_csmax=new int[10];
    for(int test=0;test<percents_csmax.length;test++){
      System.out.println("Testing Bayes"+percents_csmax[test]+"...");
      if(test!=0)
        bayes.fS_CS_Max(percents_csmax[test]);
      int count=0;
      for(int i=0;i<testset.size();i++){
        Instance data = testset.getInstance(i);
        Integer gold = (Integer) data.getTarget();
        Predict<String> pres=bayes.classify(data, Type.STRING, 3);
        String pred_label=pres.getLabel();
        String gold_label = bayes.getLabel(gold);
       
        if(pred_label.equals(gold_label)){
          count++;
        }
        else{
//          System.err.println(gold_label+"->"+pred_label+" : "+testset.getInstance(i).getTempData());
//          for(int j=0;j<3;j++)
//            System.out.println(pres.getLabel(j)+":"+pres.getScore(j));
        }
      }
      counts_csmax[test]=count;
      System.out.println("Bayes Precision("+percents_csmax[test]+"):"
      +((float)count/testset.size())+"("+count+"/"+testset.size()+")");
    }
    bayes.noFeatureSelection();
    float[] percents_ig=new float[]{1.0f,0.9f,0.8f,0.7f,0.5f,0.3f,0.2f,0.1f};
    int[] counts_ig=new int[10];
    for(int test=0;test<percents_ig.length;test++){
      System.out.println("Testing Bayes"+percents_ig[test]+"...");
      if(test!=0)
        bayes.fS_IG(percents_ig[test]);
      int count=0;
      for(int i=0;i<testset.size();i++){
        Instance data = testset.getInstance(i);
        Integer gold = (Integer) data.getTarget();
        Predict<String> pres=bayes.classify(data, Type.STRING, 3);
        String pred_label=pres.getLabel();
        String gold_label = bayes.getLabel(gold);
       
        if(pred_label.equals(gold_label)){
          count++;
        }
        else{
//          System.err.println(gold_label+"->"+pred_label+" : "+testset.getInstance(i).getTempData());
//          for(int j=0;j<3;j++)
//            System.out.println(pres.getLabel(j)+":"+pres.getScore(j));
        }
      }
      counts_ig[test]=count;
      System.out.println("Bayes Precision("+percents_csmax[test]+"):"
      +((float)count/testset.size())+"("+count+"/"+testset.size()+")");
    }
   
    System.out.println("..Testing Bayes complete!");
    for(int i=0;i<percents_cs.length;i++)
      System.out.println("Bayes Precision CS("+percents_cs[i]+"):"
    +((float)counts_cs[i]/testset.size())+"("+counts_cs[i]+"/"+testset.size()+")");
   
    for(int i=0;i<percents_csmax.length;i++)
      System.out.println("Bayes Precision CS_Max("+percents_csmax[i]+"):"
    +((float)counts_csmax[i]/testset.size())+"("+counts_csmax[i]+"/"+testset.size()+")");
   
    for(int i=0;i<percents_ig.length;i++)
      System.out.println("Bayes Precision IG("+percents_ig[i]+"):"
    +((float)counts_ig[i]/testset.size())+"("+counts_ig[i]+"/"+testset.size()+")");

  }
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
  private void init(String[][][]stringTag,String str) throws Exception{
    ts = new TreeSet<Entity>();
    llis = new LinkedList<Instance>();
    fc = new FormChanger();
    test = new InstanceSet(cl.getPipe());
    test.loadThruPipes(new AR_Reader(stringTag,str));
    for(int i=0;i<test.size();i++){
      String ss = cl.getStringLabel(test.getInstance(i));
      if(ss.equals("1")){
        llis.add(test.getInstance(i));
View Full Code Here
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
    Pipe fpipe = new StringArray2SV(factory, true);
    //构造转换器组
    Pipe pipe = new SeriesPipes(new Pipe[]{lpipe,fpipe});
   
    //构建训练集
    train = new InstanceSet(pipe, factory);
    SimpleFileReader reader = new SimpleFileReader (path,true);
    train.loadThruStagePipes(reader);
    al.setStopIncrement(true);
   
    //构建测试集
    test = new InstanceSet(pipe, factory);   
    reader = new SimpleFileReader (path,true);
    test.loadThruStagePipes(reader); 

    System.out.println("Train Number: " + train.size());
    System.out.println("Test Number: " + test.size());
View Full Code Here
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
    Pipe pipe = new SeriesPipes(new Pipe[] { new Target2Label(labels), featurePipe });


    System.out.print("读入训练数据 ...");
    InstanceSet trainSet = new InstanceSet(pipe, factory);

    // 训练集
    trainSet.loadThruStagePipes(new SequenceReader(train, true, "utf8"));
    System.out.println("训练样本个数 " + trainSet.size());
    System.out.println("标签个数: " + labels.size()); //
    System.out.println("特征个数" + features.size());

    // 冻结特征集
    features.setStopIncrement(true);
    labels.setStopIncrement(true);


    // viterbi解码
    HammingLoss loss = new HammingLoss();
    Inferencer inference = new LinearViterbi(templets, labels.size());
    Update update = new LinearViterbiPAUpdate((LinearViterbi) inference, loss);


    OnlineTrainer trainer = new OnlineTrainer(inference, update, loss,
        features.size(), 50,0.1f);

    Linear cl = trainer.train(trainSet);


    // test data没有标注
    Pipe tpipe = featurePipe;
    // 测试集
    InstanceSet testSet = new InstanceSet(tpipe);

    testSet.loadThruPipes(new SequenceReader(testfile, false, "utf8"));
    System.out.println("测试样本个数: " + testSet.size()); //
    String[][] labelsSet = new String[testSet.size()][];
    for (int i = 0; i < testSet.size(); i++) {
      Instance carrier = testSet.get(i);
      int[] pred = (int[]) cl.classify(carrier).getLabel(0);
      labelsSet[i] = labels.lookupString(pred);
    }
   
    String s = SimpleFormatter.format(testSet, labelsSet);
View Full Code Here
TOP

Related Classes of org.fnlp.ml.types.InstanceSet

  • org.fnlp.app.tc.TextClassifier
  • org.fnlp.demo.ml.HierClassifierUsage1
  • org.fnlp.demo.ml.HierClassifierUsage2
  • org.fnlp.demo.ml.SequenceLabeling
  • org.fnlp.demo.ml.SimpleClassifier2
  • org.fnlp.demo.nlp.tc.TextClassificationBasedOnBayes
  • org.fnlp.demo.nlp.tc.TextClassificationBasedOnBayes2
  • org.fnlp.demo.nlp.tc.TextClassificationBasedOnKNN
  • org.fnlp.demo.nlp.tc.TextClassificationCustom
  • org.fnlp.demo.nlp.tc.TextClassificationCustom1
    • 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.