Package de.lmu.ifi.dbs.elki.math.linearalgebra

Examples of de.lmu.ifi.dbs.elki.math.linearalgebra.EigenPair

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
    int contrastAtMax = eigenPairs.size() - 1;
    // find the last eigenvector that is considered 'strong' by the weak rule
    // applied to the remaining vectors only
    double eigenValueSum = eigenPairs.getEigenPair(eigenPairs.size() - 1).getEigenvalue();
    for(int i = eigenPairs.size() - 2; i >= 0; i--) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      eigenValueSum += eigenPair.getEigenvalue();
      double needEigenvalue = eigenValueSum / (eigenPairs.size() - i) * ralpha;
      if(eigenPair.getEigenvalue() >= needEigenvalue) {
        contrastAtMax = i;
        break;
      }
    }

    for(int i = 0; i <= contrastAtMax /* && i < eigenPairs.size() */; i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      strongEigenPairs.add(eigenPair);
    }
    for(int i = contrastAtMax + 1; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      weakEigenPairs.add(eigenPair);
    }

    return new FilteredEigenPairs(weakEigenPairs, strongEigenPairs);
  }
View Full Code Here
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
    int contrastAtMax = eigenPairs.size() - 1;
    // find the last eigenvector that is considered 'strong' by the weak rule
    // applied to the remaining vectors only
    double eigenValueSum = eigenPairs.getEigenPair(eigenPairs.size() - 1).getEigenvalue();
    for(int i = eigenPairs.size() - 2; i >= 0; i--) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      eigenValueSum += eigenPair.getEigenvalue();
      double needEigenvalue = eigenValueSum / (eigenPairs.size() - i) * ralpha;
      if(eigenPair.getEigenvalue() >= needEigenvalue) {
        contrastAtMax = i;
        break;
      }
    }

    for(int i = 0; i <= contrastAtMax /* && i < eigenPairs.size() */; i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      strongEigenPairs.add(eigenPair);
    }
    for(int i = contrastAtMax + 1; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      weakEigenPairs.add(eigenPair);
    }

    return new FilteredEigenPairs(weakEigenPairs, strongEigenPairs);
  }
View Full Code Here
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
    List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();

    // determine sum of eigenvalues
    double totalSum = 0;
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      totalSum += eigenPair.getEigenvalue();
    }
    if(logger.isDebugging()) {
      msg.append("\ntotalSum = ").append(totalSum);
    }

    // determine strong and weak eigenpairs
    double currSum = 0;
    boolean found = false;
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      currSum += eigenPair.getEigenvalue();
      if(currSum / totalSum >= alpha) {
        if(!found) {
          found = true;
          strongEigenPairs.add(eigenPair);
        }
View Full Code Here
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
      limit = delta;
    }
    else {
      double max = Double.NEGATIVE_INFINITY;
      for(int i = 0; i < eigenPairs.size(); i++) {
        EigenPair eigenPair = eigenPairs.getEigenPair(i);
        double eigenValue = Math.abs(eigenPair.getEigenvalue());
        if(max < eigenValue) {
          max = eigenValue;
        }
      }
      limit = max * delta;
    }
    if(logger.isDebugging()) {
      msg.append("\nlimit = ").append(limit);
    }

    // init strong and weak eigenpairs
    List<EigenPair> strongEigenPairs = new ArrayList<EigenPair>();
    List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();

    // determine strong and weak eigenpairs
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      double eigenValue = Math.abs(eigenPair.getEigenvalue());
      if(eigenValue >= limit) {
        strongEigenPairs.add(eigenPair);
      }
      else {
        weakEigenPairs.add(eigenPair);
View Full Code Here
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
    int contrastMaximum = eigenPairs.size() - 1;
    double maxContrast = 0.0;
    // calc the eigenvalue sum.
    double eigenValueSum = 0.0;
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      eigenValueSum += eigenPair.getEigenvalue();
    }
    double weakEigenvalue = eigenValueSum / eigenPairs.size() * walpha;
    // now find the maximum contrast.
    double currSum = eigenPairs.getEigenPair(eigenPairs.size() - 1).getEigenvalue();
    for(int i = eigenPairs.size() - 2; i >= 0; i--) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      currSum += eigenPair.getEigenvalue();
      // weak?
      if(eigenPair.getEigenvalue() < weakEigenvalue) {
        continue;
      }
      double contrast = eigenPair.getEigenvalue() / (currSum / (eigenPairs.size() - i));
      if(contrast > maxContrast) {
        maxContrast = contrast;
        contrastMaximum = i;
      }
    }

    for(int i = 0; i <= contrastMaximum /* && i < eigenPairs.size() */; i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      strongEigenPairs.add(eigenPair);
    }
    for(int i = contrastMaximum + 1; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      weakEigenPairs.add(eigenPair);
    }

    return new FilteredEigenPairs(weakEigenPairs, strongEigenPairs);
  }
View Full Code Here
61
62
63
64
65
66
67
68
69
70
71
    // initialize strong and weak eigenpairs
    // all normalized eigenpairs are regarded as strong
    final List<EigenPair> strongEigenPairs = new ArrayList<EigenPair>();
    final List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();
    for(int i = 0; i < eigenPairs.size(); i++) {
      final EigenPair eigenPair = eigenPairs.getEigenPair(i);
      normalizeEigenPair(eigenPair);
      strongEigenPairs.add(eigenPair);
    }
    if(logger.isDebugging()) {
      final StringBuffer msg = new StringBuffer();
View Full Code Here
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
    List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();

    // determine sum of eigenvalues
    double totalSum = 0;
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      totalSum += eigenPair.getEigenvalue();
    }
    double expectEigenvalue = totalSum / eigenPairs.size() * walpha;

    // determine strong and weak eigenpairs
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      if(eigenPair.getEigenvalue() > expectEigenvalue) {
        strongEigenPairs.add(eigenPair);
      }
      else {
        weakEigenPairs.add(eigenPair);
      }
View Full Code Here
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
    List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();

    // determine sum of eigenvalues
    double totalSum = 0;
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      totalSum += eigenPair.getEigenvalue();
    }
    double expectedVariance = totalSum / eigenPairs.size() * walpha;

    // determine strong and weak eigenpairs
    double currSum = 0;
    boolean found = false;
    int i;
    for(i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      // weak Eigenvector?
      if(eigenPair.getEigenvalue() < expectedVariance) {
        break;
      }
      currSum += eigenPair.getEigenvalue();
      // calculate progressive alpha level
      double alpha = 1.0 - (1.0 - palpha) * (1.0 - (i + 1) / eigenPairs.size());
      if(currSum / totalSum >= alpha || i == eigenPairs.size() - 1) {
        found = true;
        strongEigenPairs.add(eigenPair);
        break;
      }
    }
    // if we didn't hit our alpha level, we consider all vectors to be weak!
    if(!found) {
      assert (weakEigenPairs.size() == 0);
      weakEigenPairs = strongEigenPairs;
      strongEigenPairs = new ArrayList<EigenPair>();
    }
    for(; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      weakEigenPairs.add(eigenPair);
    }

    // the code using this method doesn't expect an empty strong set,
    // if we didn't find any strong ones, we make all vectors strong
View Full Code Here
85
86
87
88
89
90
91
92
93
94
95
    List<EigenPair> strongEigenPairs = new ArrayList<EigenPair>();
    List<EigenPair> weakEigenPairs = new ArrayList<EigenPair>();

    // determine strong and weak eigenpairs
    for(int i = 0; i < eigenPairs.size(); i++) {
      EigenPair eigenPair = eigenPairs.getEigenPair(i);
      if(i < n) {
        strongEigenPairs.add(eigenPair);
      }
      else {
        weakEigenPairs.add(eigenPair);
View Full Code Here
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
      List<EigenPair> strongEigenPairs = filteredEigenPairs.getStrongEigenPairs();
      strongEigenvalues = new double[strongEigenPairs.size()];
      strongEigenvectors = new Matrix(dim, strongEigenPairs.size());
      int i = 0;
      for(Iterator<EigenPair> it = strongEigenPairs.iterator(); it.hasNext(); i++) {
        EigenPair eigenPair = it.next();
        strongEigenvalues[i] = eigenPair.getEigenvalue();
        strongEigenvectors.setCol(i, eigenPair.getEigenvector());
        sumStrongEigenvalues += strongEigenvalues[i];
      }
    }

    {// weak eigenpairs
      List<EigenPair> weakEigenPairs = filteredEigenPairs.getWeakEigenPairs();
      weakEigenvalues = new double[weakEigenPairs.size()];
      weakEigenvectors = new Matrix(dim, weakEigenPairs.size());
      int i = 0;
      for(Iterator<EigenPair> it = weakEigenPairs.iterator(); it.hasNext(); i++) {
        EigenPair eigenPair = it.next();
        weakEigenvalues[i] = eigenPair.getEigenvalue();
        weakEigenvectors.setCol(i, eigenPair.getEigenvector());
        sumWeakEigenvalues += weakEigenvalues[i];
      }
    }
    explainedVariance = sumStrongEigenvalues / (sumStrongEigenvalues + sumWeakEigenvalues);
    int localdim = strongEigenvalues.length;
View Full Code Here
TOP

Related Classes of de.lmu.ifi.dbs.elki.math.linearalgebra.EigenPair

  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.FirstNEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.LimitEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.NormalizingEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.PCAFilteredResult
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.PercentageEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.ProgressiveEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.RelativeEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.SignificantEigenPairFilter
  • de.lmu.ifi.dbs.elki.math.linearalgebra.pca.WeakEigenPairFilter
  • de.lmu.ifi.dbs.elki.visualization.visualizers.scatterplot.cluster.EMClusterVisualization
    • 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.