Example usage for org.apache.mahout.cf.taste.impl.common FastIDSet clear

Introduction

In this page you can find the example usage for org.apache.mahout.cf.taste.impl.common FastIDSet clear.

Prototype

public void clear() 

Source Link

Usage

From source file:com.predictionmarketing.itemrecommend.SUSSGDFactorizer.java

License:Apache License

@Override
public Factorization factorize() throws TasteException {
    prepareTraining();/*from  w  w w. ja v a2s .  c  o  m*/
    double currentLearningRate = learningRate;
    HashSet<Long> DMusers = new HashSet<Long>();
    LongPrimitiveIterator ud = dataModel.getUserIDs();
    while (ud.hasNext()) {
        DMusers.add(ud.nextLong());
    }
    for (int it = 0; it < numIterations; it++) {
        FastIDSet friends_idset = new FastIDSet();
        double Suvpv_array[];
        // System.out.println("imawa no."+it);

        for (long userID : DMusers) {
            //useriteration
            if (SocialIDset.contains(userID)) {
                friends_idset = SocialdataModel.getItemIDsFromUser(userID);
            }
            Suvpv_array = generateSuvpv(userID, friends_idset);
            double len = vectorlength(Suvpv_array);
            assert len < 1.01;
            assert len > 0.99;
            int friends_amount = (testuser == userID) ? friends_idset.size() - 1 : friends_idset.size(); //testuser friends?-1 ()

            double mother = (double) friends_amount / 250 + 1;
            double alpha = (double) 1 - (1 / mother);
            //                double alpha=0.1;//(double)1-(1/mother);   //      social degree
            //                double alpha=0;
            assert alpha < 0.5;
            updateParameters(userID, currentLearningRate, Suvpv_array, alpha);
            friends_idset.clear();
        }
        currentLearningRate *= learningRateDecay;
    } //end learning iteration
      //System.out.println("Total records"+Total_records/10);
      //aggregate writer feature  produce a new user-write vectors replace itemVectors
    double[][] writerVectors = new double[WriterdataModel.getNumUsers()][numFeatures];
    LongPrimitiveIterator it = WriterdataModel.getUserIDs();
    while (it.hasNext()) {
        long userid = it.nextLong();
        int userIndex = userIndex(userid);
        for (int i = 0; i < numFeatures; i++) {
            writerVectors[userIndex][i] = 0.0;
        }
        int post_count = 0;
        for (long item : WriterdataModel.getItemIDsFromUser(userid)) {
            int itemIndex = itemIndex(item);
            if (itemIndex < dataModel.getNumItems()) {
                post_count++;
                for (int i = 0; i < numFeatures; i++) {
                    writerVectors[userIndex][i] += itemVectors[itemIndex][i];
                }
            }
        }
        if (post_count > 0) {
            //unit vector
            writerVectors[userIndex] = unitvectorize(writerVectors[userIndex]);
            writerVectors[userIndex][USER_BIAS_INDEX] /= post_count;
            writerVectors[userIndex][ITEM_BIAS_INDEX] /= post_count;
        }
    }
    int n = (int) (Math.random() * dataModel.getNumUsers());
    //System.out.println("writer"+n+" user bias:"+writerVectors[n][USER_BIAS_INDEX]);
    // System.out.println("writer"+n+"item bias:"+writerVectors[n][ITEM_BIAS_INDEX]);
    assert writerVectors[n][USER_BIAS_INDEX] > 0.99;
    assert writerVectors[n][USER_BIAS_INDEX] < 1.01;

    return createFactorization(userVectors, writerVectors);//(userVectors, itemVectors);
}

---

From source file:net.myrrix.common.collection.FastIDSetTest.java

License:Apache License

@Test
public void testClear() {
    FastIDSet set = new FastIDSet();
    set.add(1);/*from  w  w w. j a  va  2  s.  c om*/
    set.clear();
    assertEquals(0, set.size());
    assertTrue(set.isEmpty());
    assertFalse(set.contains(1));
}

---