// Copyright (C) 2011 Zeno Gantner, Chris Newell
//
// This file is part of MyMediaLite.
//
// MyMediaLite is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// MyMediaLite 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 MyMediaLite. If not, see <http://www.gnu.org/licenses/>.
package org.mymedialite.ratingprediction;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.PrintWriter;
import java.util.ArrayList;
import java.util.List;
import org.mymedialite.datatype.SkewSymmetricSparseMatrix;
import org.mymedialite.datatype.SymmetricSparseMatrix;
import org.mymedialite.io.IMatrixExtensions;
import org.mymedialite.io.Model;
import org.mymedialite.io.VectorExtensions;
/**
* Bi-polar frequency-weighted Slope-One rating prediction.
*
* Literature:
* Daniel Lemire, Anna Maclachlan:
* Slope One Predictors for Online Rating-Based Collaborative Filtering.
* SIAM Data Mining (SDM 2005).
* http://www.daniel-lemire.com/fr/abstracts/SDM2005.html
*
* This recommender does NOT support incremental updates. They would be easy to implement, though.
* @version 2.03
*/
public class BiPolarSlopeOne extends RatingPredictor {
private static final String VERSION = "2.03";
private SkewSymmetricSparseMatrix diff_matrix_like;
private SymmetricSparseMatrix<Integer> freq_matrix_like;
private SkewSymmetricSparseMatrix diff_matrix_dislike;
private SymmetricSparseMatrix<Integer> freq_matrix_dislike;
private double global_average;
private List<Double> user_average;
/**
*
*/
public boolean canPredict(int user_id, int item_id) {
if (user_id > maxUserID || item_id > maxItemID)
return false;
for (int index : ratings.byUser().get(user_id)) {
if (freq_matrix_like.get(item_id, ratings.items().get(index)) != 0)
return true;
if (freq_matrix_dislike.get(item_id, ratings.items().get(index)) != 0)
return true;
}
return false;
}
/**
*
*/
public double predict(int user_id, int item_id) {
if (item_id > maxItemID || user_id > maxUserID)
return global_average;
double prediction = 0.0;
int frequencies = 0;
for (int index : ratings.byUser().get(user_id)) {
if (ratings.get(index) > user_average.get(user_id)) {
int f = freq_matrix_like.get(item_id, ratings.items().get(index));
if (f != 0) {
prediction += ( diff_matrix_like.get(item_id, ratings.items().get(index)) + ratings.get(index) ) * f;
frequencies += f;
}
} else {
int f = freq_matrix_dislike.get(item_id, ratings.items().get(index));
if (f != 0) {
prediction += ( diff_matrix_dislike.get(item_id, ratings.items().get(index)) + ratings.get(index) ) * f;
frequencies += f;
}
}
}
if (frequencies == 0)
return global_average;
double result = prediction / frequencies;
if (result > maxRating)
return maxRating;
if (result < minRating)
return minRating;
return result;
}
/**
*/
public void train() {
initModel();
// Default value if no prediction can be made
global_average = ratings.average();
// Compute difference sums and frequencies
for (int user_id : ratings.allUsers()) {
double user_avg = 0;
for (int index : ratings.byUser().get(user_id))
user_avg += ratings.get(index);
user_avg = user_avg / ratings.byUser().get(user_id).size();
// Store for later use
user_average.set(user_id, user_avg);
for (int index : ratings.byUser().get(user_id)) {
for (int index2 : ratings.byUser().get(user_id)) {
if (ratings.get(index) > user_avg && ratings.get(index2) > user_avg) {
freq_matrix_like.set(ratings.items().get(index), ratings.items().get(index2), freq_matrix_like.get(ratings.items().get(index), ratings.items().get(index2)) + 1);
diff_matrix_like.set(ratings.items().get(index), ratings.items().get(index2), diff_matrix_like.get(ratings.items().get(index), ratings.items().get(index2)) + (float) (ratings.get(index) - ratings.get(index2)));
} else if (ratings.get(index) < user_avg && ratings.get(index2) < user_avg) {
freq_matrix_dislike.set(ratings.items().get(index), ratings.items().get(index2), freq_matrix_dislike.get(ratings.items().get(index), ratings.items().get(index2)) + 1);
diff_matrix_dislike.set(ratings.items().get(index), ratings.items().get(index2), diff_matrix_dislike.get(ratings.items().get(index), ratings.items().get(index2)) + (float) (ratings.get(index) - ratings.get(index2)));
}
}
}
}
// Compute average differences
for (int i = 0; i <= maxItemID; i++) {
for (int j : freq_matrix_like.get(i).keySet())
diff_matrix_like.set(i, j, diff_matrix_like.get(i, j) / freq_matrix_like.get(i, j));
for (int j : freq_matrix_dislike.get(i).keySet())
diff_matrix_dislike.set(i, j, diff_matrix_dislike.get(i, j) / freq_matrix_dislike.get(i, j));
}
}
/**
*
*/
public void initModel() {
// Create data structure
diff_matrix_like = new SkewSymmetricSparseMatrix(maxItemID + 1);
freq_matrix_like = new SymmetricSparseMatrix<Integer>(maxItemID + 1, 0);
diff_matrix_dislike = new SkewSymmetricSparseMatrix(maxItemID + 1);
freq_matrix_dislike = new SymmetricSparseMatrix<Integer>(maxItemID + 1, 0);
user_average = new ArrayList<Double>(maxUserID + 1);
for(int i=0; i < maxUserID + 1; i++) user_average.add(null);
}
@Override
public void loadModel(String filename) throws IOException {
BufferedReader reader = Model.getReader(filename, this.getClass());
loadModel(reader);
reader.close();
}
@Override
public void loadModel(BufferedReader reader) throws IOException {
initModel();
double global_average = Double.parseDouble(reader.readLine());
SkewSymmetricSparseMatrix diff_matrix_like = (SkewSymmetricSparseMatrix) IMatrixExtensions.readFloatMatrix(reader, this.diff_matrix_like);
SymmetricSparseMatrix<Integer> freq_matrix_like = (SymmetricSparseMatrix<Integer>) IMatrixExtensions.readIntegerMatrix(reader, this.freq_matrix_like);
SkewSymmetricSparseMatrix diff_matrix_dislike = (SkewSymmetricSparseMatrix) IMatrixExtensions.readFloatMatrix(reader, this.diff_matrix_dislike);
SymmetricSparseMatrix<Integer> freq_matrix_dislike = (SymmetricSparseMatrix<Integer>) IMatrixExtensions.readIntegerMatrix(reader, this.freq_matrix_dislike);
List<Double> user_average = VectorExtensions.readVector(reader);
reader.close();
// Assign new model
this.global_average = global_average;
this.diff_matrix_like = diff_matrix_like;
this.freq_matrix_like = freq_matrix_like;
this.diff_matrix_dislike = diff_matrix_dislike;
this.freq_matrix_dislike = freq_matrix_dislike;
this.user_average = user_average;
}
@Override
public void saveModel(String filename) throws IOException {
PrintWriter writer = Model.getWriter(filename, this.getClass(), VERSION);
saveModel(writer);
writer.flush();
writer.close();
}
@Override
public void saveModel(PrintWriter writer) throws IOException {
writer.println(Double.toString(global_average));
IMatrixExtensions.writeSparseMatrix(writer, diff_matrix_like);
IMatrixExtensions.writeSparseMatrix(writer, freq_matrix_like);
IMatrixExtensions.writeSparseMatrix(writer, diff_matrix_dislike);
IMatrixExtensions.writeSparseMatrix(writer, freq_matrix_dislike);
VectorExtensions.writeVector(writer, user_average);
}
}