Package org.apache.mahout.math

Examples of org.apache.mahout.math.MatrixSlice

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
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
    StringBuilder prefsAsText = new StringBuilder();
    String separator = "";
    Iterator<MatrixSlice> sliceIterator = preferences.iterateAll();
    while (sliceIterator.hasNext()) {
      MatrixSlice slice = sliceIterator.next();
      Iterator<Vector.Element> elementIterator = slice.vector().iterateNonZero();
      while (elementIterator.hasNext()) {
        Vector.Element e = elementIterator.next();
        if (!Double.isNaN(e.get())) {
          prefsAsText.append(separator).append(slice.index()).append(',').append(e.index()).append(',').append(e.get());
          separator = "\n";
        }
      }
    }
    logger.info("Input matrix:\n" + prefsAsText);
    writeLines(inputFile, prefsAsText.toString());

    ParallelALSFactorizationJob alsFactorization = new ParallelALSFactorizationJob();

    Configuration conf = new Configuration();
    conf.set("mapred.input.dir", inputFile.getAbsolutePath());
    conf.set("mapred.output.dir", outputDir.getAbsolutePath());
    conf.setBoolean("mapred.output.compress", false);

    alsFactorization.setConf(conf);
    int numFeatures = 3;
    int numIterations = 5;
    double lambda = 0.065;
    alsFactorization.run(new String[] { "--tempDir", tmpDir.getAbsolutePath(), "--lambda", String.valueOf(lambda),
        "--numFeatures", String.valueOf(numFeatures), "--numIterations", String.valueOf(numIterations) });

    Matrix u = MathHelper.readEntries(conf, new Path(outputDir.getAbsolutePath(), "U/part-r-00000"),
        preferences.numRows(), numFeatures);
    Matrix m = MathHelper.readEntries(conf, new Path(outputDir.getAbsolutePath(), "M/part-r-00000"),
      preferences.numCols(), numFeatures);

    RunningAverage avg = new FullRunningAverage();
    sliceIterator = preferences.iterateAll();
    while (sliceIterator.hasNext()) {
      MatrixSlice slice = sliceIterator.next();
      Iterator<Vector.Element> elementIterator = slice.vector().iterateNonZero();
      while (elementIterator.hasNext()) {
        Vector.Element e = elementIterator.next();
        if (!Double.isNaN(e.get())) {
          double pref = e.get();
          double estimate = u.getRow(slice.index()).dot(m.getRow(e.index()));
          double err = pref - estimate;
          avg.addDatum(err * err);
          logger.info("Comparing preference of user [" + slice.index() + "] towards item [" + e.index() + "], " +
              "was [" + pref + "] estimate is [" + estimate + ']');
        }
      }
    }
    double rmse = Math.sqrt(avg.getAverage());
View Full Code Here
45
46
47
48
49
50
51
52
53
54
55
56
57
58
    Iterator<MatrixSlice> mIt = m.iterateAll();
    Iterator<MatrixSlice> mttIt = mtt.iterateAll();
    Map<Integer, Vector> mMap = new HashMap<Integer,Vector>();
    Map<Integer, Vector> mttMap = new HashMap<Integer, Vector>();
    while (mIt.hasNext() && mttIt.hasNext()) {
      MatrixSlice ms = mIt.next();
      mMap.put(ms.index(), ms.vector());
      MatrixSlice mtts = mttIt.next();
      mttMap.put(mtts.index(), mtts.vector());
    }
    for(Map.Entry<Integer, Vector> entry : mMap.entrySet()) {
      Integer key = entry.getKey();
      Vector value = entry.getValue();
      if(value == null || mttMap.get(key) == null) {
View Full Code Here
337
338
339
340
341
342
343
344
345
346
347
348
          public boolean hasNext() {
            return it.hasNext();
          }
          @Override
          public VectorWritable next() {
            MatrixSlice slice = it.next();
            return new VectorWritable(slice.vector());
          }
          @Override
          public void remove() {
            it.remove();
          }
View Full Code Here
135
136
137
138
139
140
141
142
143
144
145
          new SequenceFileDirIterator<IntWritable,VectorWritable>(new Path(rowPath, "*"),
                                                                  PathType.GLOB, null, null, true, conf),
          new Function<Pair<IntWritable,VectorWritable>,MatrixSlice>() {
            @Override
            public MatrixSlice apply(Pair<IntWritable, VectorWritable> from) {
              return new MatrixSlice(from.getSecond().get(), from.getFirst().get());
            }
          });
    } catch (IOException ioe) {
      throw new IllegalStateException(ioe);
    }
View Full Code Here
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
    FileSystem fs = FileSystem.get(conf);
    SequenceFile.Writer seqWriter = new SequenceFile.Writer(fs, conf, path, IntWritable.class, VectorWritable.class);
    IntWritable iw = new IntWritable();
    int numEigensWritten = 0;
    for (Map.Entry<MatrixSlice, EigenStatus> pruneSlice : prunedEigenMeta) {
      MatrixSlice s = pruneSlice.getKey();
      EigenStatus meta = pruneSlice.getValue();
      EigenVector ev = new EigenVector(s.vector(),
                                       meta.getEigenValue(),
                                       Math.abs(1 - meta.getCosAngle()),
                                       s.index());
      log.info("appending {} to {}", ev, path);
      Writable vw = new VectorWritable(ev);
      iw.set(s.index());
      seqWriter.append(iw, vw);

      // increment the number of eigenvectors written and see if we've
      // reached our specified limit, or if we wish to write all eigenvectors
      // (latter is built-in, since numEigensWritten will always be > 0
View Full Code Here
191
192
193
194
195
196
197
198
199
200
201
202
203
    try {
      IntWritable iw = new IntWritable();
      int numEigensWritten = 0;
      int index = 0;
      for (Map.Entry<MatrixSlice,EigenStatus> pruneSlice : prunedEigenMeta) {
        MatrixSlice s = pruneSlice.getKey();
        EigenStatus meta = pruneSlice.getValue();
        EigenVector ev = new EigenVector(s.vector(), meta.getEigenValue(), Math.abs(1 - meta.getCosAngle()), s.index());
        // log.info("appending {} to {}", ev, path);
        Writable vw = new VectorWritable(ev);
        iw.set(index++);
        seqWriter.append(iw, vw);
View Full Code Here
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
    SVDRecommender svdRecommender = new SVDRecommender(dataModel, factorizer);

    RunningAverage avg = new FullRunningAverage();
    Iterator<MatrixSlice> sliceIterator = preferences.iterateAll();
    while (sliceIterator.hasNext()) {
      MatrixSlice slice = sliceIterator.next();
      for (Vector.Element e : slice.vector().all()) {

        long userID = slice.index() + 1;
        long itemID = e.index() + 1;

        if (!Double.isNaN(e.get())) {
          double pref = e.get();
          double estimate = svdRecommender.estimatePreference(userID, itemID);

          double confidence = 1 + alpha * observations.getQuick(slice.index(), e.index());
          double err = confidence * (pref - estimate) * (pref - estimate);
          avg.addDatum(err);
          log.info("Comparing preference of user [{}] towards item [{}], was [{}] with confidence [{}] "
              + "estimate is [{}]", slice.index(), e.index(), pref, confidence, estimate);
        }
      }
    }
    double rmse = Math.sqrt(avg.getAverage());
    log.info("RMSE: {}", rmse);
View Full Code Here
52
53
54
55
56
57
58
59
60
61
62
63
64
65
    Iterator<MatrixSlice> mIt = m.iterateAll();
    Iterator<MatrixSlice> mttIt = mtt.iterateAll();
    Map<Integer, Vector> mMap = Maps.newHashMap();
    Map<Integer, Vector> mttMap = Maps.newHashMap();
    while (mIt.hasNext() && mttIt.hasNext()) {
      MatrixSlice ms = mIt.next();
      mMap.put(ms.index(), ms.vector());
      MatrixSlice mtts = mttIt.next();
      mttMap.put(mtts.index(), mtts.vector());
    }
    for (Map.Entry<Integer, Vector> entry : mMap.entrySet()) {
      Integer key = entry.getKey();
      Vector value = entry.getValue();
      if (value == null || mttMap.get(key) == null) {
View Full Code Here
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
    SequenceFile.Writer seqWriter = new SequenceFile.Writer(fs, conf, path, IntWritable.class, VectorWritable.class);
    try {
      IntWritable iw = new IntWritable();
      int numEigensWritten = 0;
      for (Map.Entry<MatrixSlice, EigenStatus> pruneSlice : prunedEigenMeta) {
        MatrixSlice s = pruneSlice.getKey();
        EigenStatus meta = pruneSlice.getValue();
        EigenVector ev = new EigenVector(s.vector(),
                                         meta.getEigenValue(),
                                         Math.abs(1 - meta.getCosAngle()),
                                         s.index());
        //log.info("appending {} to {}", ev, path);
        Writable vw = new VectorWritable(ev);
        iw.set(s.index());
        seqWriter.append(iw, vw);

        // increment the number of eigenvectors written and see if we've
        // reached our specified limit, or if we wish to write all eigenvectors
        // (latter is built-in, since numEigensWritten will always be > 0
View Full Code Here
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
    Iterator<MatrixSlice> docIterator = matrix.iterator();
    Iterator<MatrixSlice> docTopicIterator = docTopicCounts.iterator();
    double perplexity = 0;
    double matrixNorm = 0;
    while (docIterator.hasNext() && docTopicIterator.hasNext()) {
      MatrixSlice docSlice = docIterator.next();
      MatrixSlice topicSlice = docTopicIterator.next();
      int docId = docSlice.index();
      Vector document = docSlice.vector();
      Vector topicDist = topicSlice.vector();
      if (testFraction == 0 || docId % (1/testFraction) == 0) {
        trainSync(document, topicDist, false, 10);
        perplexity += readModel.perplexity(document, topicDist);
        matrixNorm += document.norm(1);
      }
View Full Code Here
TOP

Related Classes of org.apache.mahout.math.MatrixSlice

  • org.apache.mahout.cf.taste.hadoop.als.ParallelALSFactorizationJobTest
  • org.apache.mahout.cf.taste.impl.recommender.svd.ALSWRFactorizerTest
  • org.apache.mahout.clustering.lda.cvb.ModelTrainer
  • org.apache.mahout.math.hadoop.decomposer.EigenVerificationJob
  • org.apache.mahout.math.hadoop.DistributedRowMatrix
  • org.apache.mahout.math.hadoop.DistributedRowMatrix$DistributedMatrixIterator
  • org.apache.mahout.math.hadoop.TestDistributedRowMatrix
    • 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.