Examples with structures._SparseFeature structures._SparseFeature used on opensource projects

Example 71 with structures._SparseFeature

use of structures._SparseFeature in project IR_Base by Linda-sunshine.

the class pLSA method calculate_log_likelihood.

/*likelihod calculation */
/* M is number of doc
	 * N is number of word in corpus
	 */
/* p(w,d) = sum_1_M sum_1_N count(d_i, w_j) * log[ lambda*p(w|theta_B) + [lambda * sum_1_k (p(w|z) * p(z|d)) */
// NOTE: cannot be used for unseen documents!
@Override
protected double calculate_log_likelihood(_Doc d) {
    double logLikelihood = docThetaLikelihood(d), prob;
    for (_SparseFeature fv : d.getSparse()) {
        int j = fv.getIndex();
        prob = 0.0;
        for (// \sum_z p(w|z,\theta)p(z|d)
        int k = 0; // \sum_z p(w|z,\theta)p(z|d)
        k < this.number_of_topics; // \sum_z p(w|z,\theta)p(z|d)
        k++) prob += d.m_topics[k] * topic_term_probabilty[k][j];
        // (1-\lambda)p(w|d) * \lambda p(w|theta_b)
        prob = prob * (1 - m_lambda) + this.background_probability[j] * m_lambda;
        logLikelihood += fv.getValue() * Math.log(prob);
    }
    return logLikelihood;
}

Example 72 with structures._SparseFeature

use of structures._SparseFeature in project IR_Base by Linda-sunshine.

the class twoTopic method calculate_log_likelihood.

protected double calculate_log_likelihood(_Doc d) {
    double logLikelihood = 0.0;
    for (_SparseFeature fv : d.getSparse()) {
        int wid = fv.getIndex();
        logLikelihood += fv.getValue() * Math.log(m_lambda * background_probability[wid] + (1 - m_lambda) * m_theta[wid]);
    }
    return logLikelihood;
}

Example 73 with structures._SparseFeature

use of structures._SparseFeature in project IR_Base by Linda-sunshine.

the class twoTopic method calculate_E_step.

@Override
public double calculate_E_step(_Doc d) {
    for (_SparseFeature fv : d.getSparse()) {
        int wid = fv.getIndex();
        m_sstat[wid] = (1 - m_lambda) * m_theta[wid];
        // compute the expectation
        m_sstat[wid] = fv.getValue() * m_sstat[wid] / (m_sstat[wid] + m_lambda * background_probability[wid]);
    }
    return calculate_log_likelihood(d);
}

Example 74 with structures._SparseFeature

use of structures._SparseFeature in project IR_Base by Linda-sunshine.

the class Utils method dotProduct.

// Calculate the similarity between two sparse vectors.
public static double dotProduct(_SparseFeature[] spVct1, _SparseFeature[] spVct2) {
    if (spVct1 == null || spVct2 == null)
        // What is the minimal value of similarity?
        return 0;
    double similarity = 0;
    int p1 = 0, p2 = 0;
    while (p1 < spVct1.length && p2 < spVct2.length) {
        _SparseFeature t1 = spVct1[p1];
        _SparseFeature t2 = spVct2[p2];
        if (t1.getIndex() == t2.getIndex()) {
            similarity += t1.getValue() * t2.getValue();
            p1++;
            p2++;
        } else if (t1.getIndex() > t2.getIndex())
            p2++;
        else
            p1++;
    }
    return similarity;
}

Example 75 with structures._SparseFeature

use of structures._SparseFeature in project IR_Base by Linda-sunshine.

the class Utils method createLibLinearFV.

public static Feature[] createLibLinearFV(HashMap<Integer, Double> spVct) {
    Feature[] node = new Feature[spVct.size()];
    int fid = 0;
    for (_SparseFeature fv : createSpVct(spVct)) // svm's feature index starts from 1
    node[fid++] = new FeatureNode(1 + fv.getIndex(), fv.getValue());
    return node;
}