/*
* LensKit, an open source recommender systems toolkit.
* Copyright 2010-2014 LensKit Contributors. See CONTRIBUTORS.md.
* Work on LensKit has been funded by the National Science Foundation under
* grants IIS 05-34939, 08-08692, 08-12148, and 10-17697.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
* FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
* details.
*
* You should have received a copy of the GNU General Public License along with
* this program; if not, write to the Free Software Foundation, Inc., 51
* Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
package org.grouplens.lenskit.mf.funksvd;
import it.unimi.dsi.fastutil.longs.LongCollection;
import it.unimi.dsi.fastutil.longs.LongIterator;
import mikera.vectorz.AVector;
import org.grouplens.lenskit.ItemScorer;
import org.grouplens.lenskit.data.pref.IndexedPreference;
import org.grouplens.lenskit.data.pref.PreferenceDomain;
import org.grouplens.lenskit.data.snapshot.PreferenceSnapshot;
import org.grouplens.lenskit.vectors.MutableSparseVector;
import org.grouplens.lenskit.vectors.SparseVector;
import java.util.Collection;
/**
* Rating estimates used while training the predictor. An estimator can be constructed
* using {@link FunkSVDUpdateRule#makeEstimator(PreferenceSnapshot)}.
*
* @author <a href="http://www.grouplens.org">GroupLens Research</a>
* @since 1.1
*/
public final class TrainingEstimator {
private final Collection<IndexedPreference> ratings;
private final double[] estimates;
private final PreferenceDomain domain;
/**
* Initialize the training estimator.
*
* @param snap The preference snapshot.
* @param baseline The baseline predictor.
* @param dom The preference domain (for clamping).
*/
TrainingEstimator(PreferenceSnapshot snap, ItemScorer baseline, PreferenceDomain dom) {
ratings = snap.getRatings();
domain = dom;
estimates = new double[ratings.size()];
final LongCollection userIds = snap.getUserIds();
LongIterator userIter = userIds.iterator();
while (userIter.hasNext()) {
long uid = userIter.nextLong();
SparseVector rvector = snap.userRatingVector(uid);
MutableSparseVector blpreds = MutableSparseVector.create(rvector.keySet());
baseline.score(uid, blpreds);
for (IndexedPreference r : snap.getUserRatings(uid)) {
estimates[r.getIndex()] = blpreds.get(r.getItemId());
}
}
}
/**
* Get the estimate for a preference.
* @param pref The preference.
* @return The estimate.
*/
public double get(IndexedPreference pref) {
return estimates[pref.getIndex()];
}
/**
* Update the current estimates with trained values for a new feature.
* @param ufvs The user feature values.
* @param ifvs The item feature values.
*/
public void update(AVector ufvs, AVector ifvs) {
for (IndexedPreference r : ratings) {
int idx = r.getIndex();
double est = estimates[idx];
est += ufvs.get(r.getUserIndex()) * ifvs.get(r.getItemIndex());
if (domain != null) {
est = domain.clampValue(est);
}
estimates[idx] = est;
}
}
}