List of usage examples for org.apache.mahout.math Vector plus
Vector plus(Vector x);
From source file:cn.edu.bjtu.cit.recommender.Recommender.java
License:Apache License
@SuppressWarnings("unchecked") public int run(String[] args) throws Exception { if (args.length < 2) { System.err.println();//from w w w .j a v a2 s .co m System.err.println("Usage: " + this.getClass().getName() + " [generic options] input output [profiling] [estimation] [clustersize]"); System.err.println(); printUsage(); GenericOptionsParser.printGenericCommandUsage(System.err); return 1; } OptionParser parser = new OptionParser(args); Pipeline pipeline = new MRPipeline(Recommender.class, getConf()); if (parser.hasOption(CLUSTER_SIZE)) { pipeline.getConfiguration().setInt(ClusterOracle.CLUSTER_SIZE, Integer.parseInt(parser.getOption(CLUSTER_SIZE).getValue())); } if (parser.hasOption(PROFILING)) { pipeline.getConfiguration().setBoolean(Profiler.IS_PROFILE, true); this.profileFilePath = parser.getOption(PROFILING).getValue(); } if (parser.hasOption(ESTIMATION)) { estFile = parser.getOption(ESTIMATION).getValue(); est = new Estimator(estFile, clusterSize); } if (parser.hasOption(OPT_REDUCE)) { pipeline.getConfiguration().setBoolean(OPT_REDUCE, true); } if (parser.hasOption(OPT_MSCR)) { pipeline.getConfiguration().setBoolean(OPT_MSCR, true); } if (parser.hasOption(ACTIVE_THRESHOLD)) { threshold = Integer.parseInt(parser.getOption("at").getValue()); } if (parser.hasOption(TOP)) { top = Integer.parseInt(parser.getOption("top").getValue()); } profiler = new Profiler(pipeline); /* * input node */ PCollection<String> lines = pipeline.readTextFile(args[0]); if (profiler.isProfiling() && lines.getSize() > 10 * 1024 * 1024) { lines = lines.sample(0.1); } /* * S0 + GBK */ PGroupedTable<Long, Long> userWithPrefs = lines.parallelDo(new MapFn<String, Pair<Long, Long>>() { @Override public Pair<Long, Long> map(String input) { String[] split = input.split(Estimator.DELM); long userID = Long.parseLong(split[0]); long itemID = Long.parseLong(split[1]); return Pair.of(userID, itemID); } @Override public float scaleFactor() { return est.getScaleFactor("S0").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S0").recsFactor; } }, Writables.tableOf(Writables.longs(), Writables.longs())).groupByKey(est.getClusterSize()); /* * S1 */ PTable<Long, Vector> userVector = userWithPrefs .parallelDo(new MapFn<Pair<Long, Iterable<Long>>, Pair<Long, Vector>>() { @Override public Pair<Long, Vector> map(Pair<Long, Iterable<Long>> input) { Vector userVector = new RandomAccessSparseVector(Integer.MAX_VALUE, 100); for (long itemPref : input.second()) { userVector.set((int) itemPref, 1.0f); } return Pair.of(input.first(), userVector); } @Override public float scaleFactor() { return est.getScaleFactor("S1").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S1").recsFactor; } }, Writables.tableOf(Writables.longs(), Writables.vectors())); userVector = profiler.profile("S0-S1", pipeline, userVector, ProfileConverter.long_vector(), Writables.tableOf(Writables.longs(), Writables.vectors())); /* * S2 */ PTable<Long, Vector> filteredUserVector = userVector .parallelDo(new DoFn<Pair<Long, Vector>, Pair<Long, Vector>>() { @Override public void process(Pair<Long, Vector> input, Emitter<Pair<Long, Vector>> emitter) { if (input.second().getNumNondefaultElements() > threshold) { emitter.emit(input); } } @Override public float scaleFactor() { return est.getScaleFactor("S2").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S2").recsFactor; } }, Writables.tableOf(Writables.longs(), Writables.vectors())); filteredUserVector = profiler.profile("S2", pipeline, filteredUserVector, ProfileConverter.long_vector(), Writables.tableOf(Writables.longs(), Writables.vectors())); /* * S3 + GBK */ PGroupedTable<Integer, Integer> coOccurencePairs = filteredUserVector .parallelDo(new DoFn<Pair<Long, Vector>, Pair<Integer, Integer>>() { @Override public void process(Pair<Long, Vector> input, Emitter<Pair<Integer, Integer>> emitter) { Iterator<Vector.Element> it = input.second().iterateNonZero(); while (it.hasNext()) { int index1 = it.next().index(); Iterator<Vector.Element> it2 = input.second().iterateNonZero(); while (it2.hasNext()) { int index2 = it2.next().index(); emitter.emit(Pair.of(index1, index2)); } } } @Override public float scaleFactor() { float size = est.getScaleFactor("S3").sizeFactor; return size; } @Override public float scaleFactorByRecord() { float recs = est.getScaleFactor("S3").recsFactor; return recs; } }, Writables.tableOf(Writables.ints(), Writables.ints())).groupByKey(est.getClusterSize()); /* * S4 */ PTable<Integer, Vector> coOccurenceVector = coOccurencePairs .parallelDo(new MapFn<Pair<Integer, Iterable<Integer>>, Pair<Integer, Vector>>() { @Override public Pair<Integer, Vector> map(Pair<Integer, Iterable<Integer>> input) { Vector cooccurrenceRow = new RandomAccessSparseVector(Integer.MAX_VALUE, 100); for (int itemIndex2 : input.second()) { cooccurrenceRow.set(itemIndex2, cooccurrenceRow.get(itemIndex2) + 1.0); } return Pair.of(input.first(), cooccurrenceRow); } @Override public float scaleFactor() { return est.getScaleFactor("S4").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S4").recsFactor; } }, Writables.tableOf(Writables.ints(), Writables.vectors())); coOccurenceVector = profiler.profile("S3-S4", pipeline, coOccurenceVector, ProfileConverter.int_vector(), Writables.tableOf(Writables.ints(), Writables.vectors())); /* * S5 Wrapping co-occurrence columns */ PTable<Integer, VectorOrPref> wrappedCooccurrence = coOccurenceVector .parallelDo(new MapFn<Pair<Integer, Vector>, Pair<Integer, VectorOrPref>>() { @Override public Pair<Integer, VectorOrPref> map(Pair<Integer, Vector> input) { return Pair.of(input.first(), new VectorOrPref(input.second())); } @Override public float scaleFactor() { return est.getScaleFactor("S5").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S5").recsFactor; } }, Writables.tableOf(Writables.ints(), VectorOrPref.vectorOrPrefs())); wrappedCooccurrence = profiler.profile("S5", pipeline, wrappedCooccurrence, ProfileConverter.int_vopv(), Writables.tableOf(Writables.ints(), VectorOrPref.vectorOrPrefs())); /* * S6 Splitting user vectors */ PTable<Integer, VectorOrPref> userVectorSplit = filteredUserVector .parallelDo(new DoFn<Pair<Long, Vector>, Pair<Integer, VectorOrPref>>() { @Override public void process(Pair<Long, Vector> input, Emitter<Pair<Integer, VectorOrPref>> emitter) { long userID = input.first(); Vector userVector = input.second(); Iterator<Vector.Element> it = userVector.iterateNonZero(); while (it.hasNext()) { Vector.Element e = it.next(); int itemIndex = e.index(); float preferenceValue = (float) e.get(); emitter.emit(Pair.of(itemIndex, new VectorOrPref(userID, preferenceValue))); } } @Override public float scaleFactor() { return est.getScaleFactor("S6").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S6").recsFactor; } }, Writables.tableOf(Writables.ints(), VectorOrPref.vectorOrPrefs())); userVectorSplit = profiler.profile("S6", pipeline, userVectorSplit, ProfileConverter.int_vopp(), Writables.tableOf(Writables.ints(), VectorOrPref.vectorOrPrefs())); /* * S7 Combine VectorOrPrefs */ PTable<Integer, VectorAndPrefs> combinedVectorOrPref = wrappedCooccurrence.union(userVectorSplit) .groupByKey(est.getClusterSize()) .parallelDo(new DoFn<Pair<Integer, Iterable<VectorOrPref>>, Pair<Integer, VectorAndPrefs>>() { @Override public void process(Pair<Integer, Iterable<VectorOrPref>> input, Emitter<Pair<Integer, VectorAndPrefs>> emitter) { Vector vector = null; List<Long> userIDs = Lists.newArrayList(); List<Float> values = Lists.newArrayList(); for (VectorOrPref vop : input.second()) { if (vector == null) { vector = vop.getVector(); } long userID = vop.getUserID(); if (userID != Long.MIN_VALUE) { userIDs.add(vop.getUserID()); } float value = vop.getValue(); if (!Float.isNaN(value)) { values.add(vop.getValue()); } } emitter.emit(Pair.of(input.first(), new VectorAndPrefs(vector, userIDs, values))); } @Override public float scaleFactor() { return est.getScaleFactor("S7").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S7").recsFactor; } }, Writables.tableOf(Writables.ints(), VectorAndPrefs.vectorAndPrefs())); combinedVectorOrPref = profiler.profile("S5+S6-S7", pipeline, combinedVectorOrPref, ProfileConverter.int_vap(), Writables.tableOf(Writables.ints(), VectorAndPrefs.vectorAndPrefs())); /* * S8 Computing partial recommendation vectors */ PTable<Long, Vector> partialMultiply = combinedVectorOrPref .parallelDo(new DoFn<Pair<Integer, VectorAndPrefs>, Pair<Long, Vector>>() { @Override public void process(Pair<Integer, VectorAndPrefs> input, Emitter<Pair<Long, Vector>> emitter) { Vector cooccurrenceColumn = input.second().getVector(); List<Long> userIDs = input.second().getUserIDs(); List<Float> prefValues = input.second().getValues(); for (int i = 0; i < userIDs.size(); i++) { long userID = userIDs.get(i); if (userID != Long.MIN_VALUE) { float prefValue = prefValues.get(i); Vector partialProduct = cooccurrenceColumn.times(prefValue); emitter.emit(Pair.of(userID, partialProduct)); } } } @Override public float scaleFactor() { return est.getScaleFactor("S8").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S8").recsFactor; } }, Writables.tableOf(Writables.longs(), Writables.vectors())).groupByKey(est.getClusterSize()) .combineValues(new CombineFn<Long, Vector>() { @Override public void process(Pair<Long, Iterable<Vector>> input, Emitter<Pair<Long, Vector>> emitter) { Vector partial = null; for (Vector vector : input.second()) { partial = partial == null ? vector : partial.plus(vector); } emitter.emit(Pair.of(input.first(), partial)); } @Override public float scaleFactor() { return est.getScaleFactor("combine").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("combine").recsFactor; } }); partialMultiply = profiler.profile("S8-combine", pipeline, partialMultiply, ProfileConverter.long_vector(), Writables.tableOf(Writables.longs(), Writables.vectors())); /* * S9 Producing recommendations from vectors */ PTable<Long, RecommendedItems> recommendedItems = partialMultiply .parallelDo(new DoFn<Pair<Long, Vector>, Pair<Long, RecommendedItems>>() { @Override public void process(Pair<Long, Vector> input, Emitter<Pair<Long, RecommendedItems>> emitter) { Queue<RecommendedItem> topItems = new PriorityQueue<RecommendedItem>(11, Collections.reverseOrder(BY_PREFERENCE_VALUE)); Iterator<Vector.Element> recommendationVectorIterator = input.second().iterateNonZero(); while (recommendationVectorIterator.hasNext()) { Vector.Element element = recommendationVectorIterator.next(); int index = element.index(); float value = (float) element.get(); if (topItems.size() < top) { topItems.add(new GenericRecommendedItem(index, value)); } else if (value > topItems.peek().getValue()) { topItems.add(new GenericRecommendedItem(index, value)); topItems.poll(); } } List<RecommendedItem> recommendations = new ArrayList<RecommendedItem>(topItems.size()); recommendations.addAll(topItems); Collections.sort(recommendations, BY_PREFERENCE_VALUE); emitter.emit(Pair.of(input.first(), new RecommendedItems(recommendations))); } @Override public float scaleFactor() { return est.getScaleFactor("S9").sizeFactor; } @Override public float scaleFactorByRecord() { return est.getScaleFactor("S9").recsFactor; } }, Writables.tableOf(Writables.longs(), RecommendedItems.recommendedItems())); recommendedItems = profiler.profile("S9", pipeline, recommendedItems, ProfileConverter.long_ri(), Writables.tableOf(Writables.longs(), RecommendedItems.recommendedItems())); /* * Profiling */ if (profiler.isProfiling()) { profiler.writeResultToFile(profileFilePath); profiler.cleanup(pipeline.getConfiguration()); return 0; } /* * asText */ pipeline.writeTextFile(recommendedItems, args[1]); PipelineResult result = pipeline.done(); return result.succeeded() ? 0 : 1; }
From source file:com.cloudera.science.ml.kmeans.core.KMeans.java
License:Open Source License
/** * Compute the {@code Vector} that is the centroid of the given weighted points. * //from w ww. j a v a 2s .co m * @param points The weighted points * @return The centroid of the weighted points */ public <V extends Vector> Vector centroid(Collection<Weighted<V>> points) { Vector center = null; long sz = 0; for (Weighted<V> v : points) { Vector weighted = v.thing().times(v.weight()); if (center == null) { center = weighted; } else { center = center.plus(weighted); } sz += v.weight(); } return center.divide(sz); }
From source file:com.cloudera.science.ml.kmeans.core.LloydsUpdateStrategy.java
License:Open Source License
/** * Compute the {@code Vector} that is the centroid of the given weighted points. * /* w ww.ja va2 s. c o m*/ * @param points The weighted points * @return The centroid of the weighted points */ public <V extends Vector> Vector centroid(Collection<Weighted<V>> points) { Vector center = null; double sz = 0.0; for (Weighted<V> v : points) { Vector weighted = v.thing().times(v.weight()); if (center == null) { center = weighted; } else { center = center.plus(weighted); } sz += v.weight(); } return center.divide(sz); }
From source file:com.cloudera.science.ml.kmeans.core.MiniBatchUpdateStrategy.java
License:Open Source License
@Override public <V extends Vector> Centers update(List<Weighted<V>> points, Centers centers) { int[] perCenterStepCounts = new int[centers.size()]; WeightedSampler<V> sampler = new WeightedSampler<V>(points, random); for (int iter = 0; iter < numIterations; iter++) { // Compute closest cent for each mini-batch List<List<V>> centerAssignments = Lists.newArrayList(); for (int i = 0; i < centers.size(); i++) { centerAssignments.add(Lists.<V>newArrayList()); }//from www. j a va 2s. c o m for (int i = 0; i < miniBatchSize; i++) { V sample = sampler.sample(); int closestId = centers.indexOfClosest(sample); centerAssignments.get(closestId).add(sample); } // Apply the mini-batch List<Vector> nextCenters = Lists.newArrayList(); for (int i = 0; i < centerAssignments.size(); i++) { Vector currentCenter = centers.get(i); for (int j = 0; j < centerAssignments.get(i).size(); j++) { double eta = 1.0 / (++perCenterStepCounts[i] + 1.0); currentCenter = currentCenter.times(1.0 - eta); currentCenter = currentCenter.plus(centerAssignments.get(i).get(j).times(eta)); } nextCenters.add(currentCenter); } centers = new Centers(nextCenters); } return centers; }
From source file:nl.gridline.zieook.inx.movielens.AggregateAndRecommendReducer.java
License:Apache License
private void reduceBooleanData(VarLongWritable userID, Iterable<PrefAndSimilarityColumnWritable> values, Context context) throws IOException, InterruptedException { /*/*from ww w . j a v a2 s .c o m*/ * having boolean data, each estimated preference can only be 1, * however we can't use this to rank the recommended items, * so we use the sum of similarities for that. */ Vector predictionVector = null; for (PrefAndSimilarityColumnWritable prefAndSimilarityColumn : values) { predictionVector = predictionVector == null ? prefAndSimilarityColumn.getSimilarityColumn() : predictionVector.plus(prefAndSimilarityColumn.getSimilarityColumn()); } writeRecommendedItems(userID, predictionVector, context); }
From source file:nl.gridline.zieook.inx.movielens.AggregateAndRecommendReducer.java
License:Apache License
private void reduceNonBooleanData(VarLongWritable userID, Iterable<PrefAndSimilarityColumnWritable> values, Context context) throws IOException, InterruptedException { /* each entry here is the sum in the numerator of the prediction formula */ Vector numerators = null; /* each entry here is the sum in the denominator of the prediction formula */ Vector denominators = null;//from w w w . ja va 2s .c o m /* each entry here is the number of similar items used in the prediction formula */ Vector numberOfSimilarItemsUsed = new RandomAccessSparseVector(Integer.MAX_VALUE, 100); for (PrefAndSimilarityColumnWritable prefAndSimilarityColumn : values) { Vector simColumn = prefAndSimilarityColumn.getSimilarityColumn(); float prefValue = prefAndSimilarityColumn.getPrefValue(); /* count the number of items used for each prediction */ Iterator<Vector.Element> usedItemsIterator = simColumn.iterateNonZero(); while (usedItemsIterator.hasNext()) { int itemIDIndex = usedItemsIterator.next().index(); numberOfSimilarItemsUsed.setQuick(itemIDIndex, numberOfSimilarItemsUsed.getQuick(itemIDIndex) + 1); } numerators = numerators == null ? prefValue == BOOLEAN_PREF_VALUE ? simColumn.clone() : simColumn.times(prefValue) : numerators.plus(prefValue == BOOLEAN_PREF_VALUE ? simColumn : simColumn.times(prefValue)); simColumn.assign(ABSOLUTE_VALUES); denominators = denominators == null ? simColumn : denominators.plus(simColumn); } if (numerators == null) { return; } Vector recommendationVector = new RandomAccessSparseVector(Integer.MAX_VALUE, 100); Iterator<Vector.Element> iterator = numerators.iterateNonZero(); while (iterator.hasNext()) { Vector.Element element = iterator.next(); int itemIDIndex = element.index(); /* preference estimations must be based on at least 2 datapoints */ if (numberOfSimilarItemsUsed.getQuick(itemIDIndex) > 1) { /* compute normalized prediction */ double prediction = element.get() / denominators.getQuick(itemIDIndex); recommendationVector.setQuick(itemIDIndex, prediction); } } writeRecommendedItems(userID, recommendationVector, context); }
From source file:org.trustedanalytics.atk.giraph.aggregators.VectorSumAggregator.java
License:Apache License
@Override public void aggregate(VectorWritable value) { Vector currentValue = getAggregatedValue().get(); if (currentValue.size() == 0) { getAggregatedValue().set(value.get()); } else if (value.get().size() > 0) { getAggregatedValue().set(currentValue.plus(value.get())); }//from w w w . j a v a 2 s.com }
From source file:org.trustedanalytics.atk.giraph.algorithms.als.AlternatingLeastSquaresComputation.java
License:Apache License
@Override public void compute(Vertex<CFVertexId, VertexData4CFWritable, EdgeData4CFWritable> vertex, Iterable<MessageData4CFWritable> messages) throws IOException { long step = getSuperstep(); if (step == 0) { initialize(vertex);/*from w w w .j a v a2 s . c om*/ vertex.voteToHalt(); return; } Vector currentValue = vertex.getValue().getVector(); double currentBias = vertex.getValue().getBias(); // update aggregators every (2 * interval) super steps if ((step % (2 * learningCurveOutputInterval)) == 0) { double errorOnTrain = 0d; double errorOnValidate = 0d; double errorOnTest = 0d; int numTrain = 0; for (MessageData4CFWritable message : messages) { EdgeType et = message.getType(); double weight = message.getWeight(); Vector vector = message.getVector(); double otherBias = message.getBias(); double predict = currentBias + otherBias + currentValue.dot(vector); double e = weight - predict; switch (et) { case TRAIN: errorOnTrain += e * e; numTrain++; break; case VALIDATE: errorOnValidate += e * e; break; case TEST: errorOnTest += e * e; break; default: throw new IllegalArgumentException("Unknown recognized edge type: " + et.toString()); } } double costOnTrain = 0d; if (numTrain > 0) { costOnTrain = errorOnTrain / numTrain + lambda * (currentBias * currentBias + currentValue.dot(currentValue)); } aggregate(SUM_TRAIN_COST, new DoubleWritable(costOnTrain)); aggregate(SUM_VALIDATE_ERROR, new DoubleWritable(errorOnValidate)); aggregate(SUM_TEST_ERROR, new DoubleWritable(errorOnTest)); } // update vertex value if (step < maxSupersteps) { // xxt records the result of x times x transpose Matrix xxt = new DenseMatrix(featureDimension, featureDimension); xxt = xxt.assign(0d); // xr records the result of x times rating Vector xr = currentValue.clone().assign(0d); int numTrain = 0; for (MessageData4CFWritable message : messages) { EdgeType et = message.getType(); if (et == EdgeType.TRAIN) { double weight = message.getWeight(); Vector vector = message.getVector(); double otherBias = message.getBias(); xxt = xxt.plus(vector.cross(vector)); xr = xr.plus(vector.times(weight - currentBias - otherBias)); numTrain++; } } xxt = xxt.plus(new DiagonalMatrix(lambda * numTrain, featureDimension)); Matrix bMatrix = new DenseMatrix(featureDimension, 1).assignColumn(0, xr); Vector value = new QRDecomposition(xxt).solve(bMatrix).viewColumn(0); vertex.getValue().setVector(value); // update vertex bias if (biasOn) { double bias = computeBias(value, messages); vertex.getValue().setBias(bias); } // send out messages for (Edge<CFVertexId, EdgeData4CFWritable> edge : vertex.getEdges()) { MessageData4CFWritable newMessage = new MessageData4CFWritable(vertex.getValue(), edge.getValue()); sendMessage(edge.getTargetVertexId(), newMessage); } } vertex.voteToHalt(); }
From source file:org.trustedanalytics.atk.giraph.algorithms.cgd.ConjugateGradientDescentComputation.java
License:Apache License
/** * Compute gradient//from w ww .j a v a2 s .c om * * @param bias of type double * @param value of type Vector * @param messages of type Iterable * @return gradient of type Vector */ private Vector computeGradient(double bias, Vector value, Iterable<MessageData4CFWritable> messages) { Vector xr = value.clone().assign(0d); int numTrain = 0; for (MessageData4CFWritable message : messages) { EdgeType et = message.getType(); if (et == EdgeType.TRAIN) { double weight = message.getWeight(); Vector vector = message.getVector(); double otherBias = message.getBias(); double predict = bias + otherBias + value.dot(vector); double e = predict - weight; xr = xr.plus(vector.times(e)); numTrain++; } } Vector gradient = value.clone().assign(0d); if (numTrain > 0) { gradient = xr.divide(numTrain).plus(value.times(lambda)); } return gradient; }
From source file:org.trustedanalytics.atk.giraph.algorithms.cgd.ConjugateGradientDescentComputation.java
License:Apache License
@Override public void compute(Vertex<CFVertexId, VertexData4CGDWritable, EdgeData4CFWritable> vertex, Iterable<MessageData4CFWritable> messages) throws IOException { long step = getSuperstep(); if (step == 0) { initialize(vertex);//from ww w . j a v a 2 s .c o m vertex.voteToHalt(); return; } Vector currentValue = vertex.getValue().getVector(); double currentBias = vertex.getValue().getBias(); // update aggregators every (2 * interval) super steps if ((step % (2 * learningCurveOutputInterval)) == 0) { double errorOnTrain = 0d; double errorOnValidate = 0d; double errorOnTest = 0d; int numTrain = 0; for (MessageData4CFWritable message : messages) { EdgeType et = message.getType(); double weight = message.getWeight(); Vector vector = message.getVector(); double otherBias = message.getBias(); double predict = currentBias + otherBias + currentValue.dot(vector); double e = weight - predict; switch (et) { case TRAIN: errorOnTrain += e * e; numTrain++; break; case VALIDATE: errorOnValidate += e * e; break; case TEST: errorOnTest += e * e; break; default: throw new IllegalArgumentException("Unknown recognized edge type: " + et.toString()); } } double costOnTrain = 0d; if (numTrain > 0) { costOnTrain = errorOnTrain / numTrain + lambda * (currentBias * currentBias + currentValue.dot(currentValue)); } aggregate(SUM_TRAIN_COST, new DoubleWritable(costOnTrain)); aggregate(SUM_VALIDATE_ERROR, new DoubleWritable(errorOnValidate)); aggregate(SUM_TEST_ERROR, new DoubleWritable(errorOnTest)); } if (step < maxSupersteps) { // implement CGD iterations Vector value0 = vertex.getValue().getVector(); Vector gradient0 = vertex.getValue().getGradient(); Vector conjugate0 = vertex.getValue().getConjugate(); double bias0 = vertex.getValue().getBias(); for (int i = 0; i < numCGDIters; i++) { double alpha = computeAlpha(gradient0, conjugate0, messages); Vector value = value0.plus(conjugate0.times(alpha)); Vector gradient = computeGradient(bias0, value, messages); double beta = computeBeta(gradient0, conjugate0, gradient); Vector conjugate = conjugate0.times(beta).minus(gradient); value0 = value; gradient0 = gradient; conjugate0 = conjugate; } // update vertex values vertex.getValue().setVector(value0); vertex.getValue().setConjugate(conjugate0); vertex.getValue().setGradient(gradient0); // update vertex bias if (biasOn) { double bias = computeBias(value0, messages); vertex.getValue().setBias(bias); } // send out messages for (Edge<CFVertexId, EdgeData4CFWritable> edge : vertex.getEdges()) { MessageData4CFWritable newMessage = new MessageData4CFWritable(vertex.getValue(), edge.getValue()); sendMessage(edge.getTargetVertexId(), newMessage); } } vertex.voteToHalt(); }