netflix.recommender
Class IncRecommender

java.lang.Object
  netflix.recommender.AbstractRecommender
      netflix.recommender.IncRecommender

public class IncRecommender
extends AbstractRecommender

Uses a serialized IncrementalSVDMovielens object to make predictions using the AbstractRecommender framework. This recommender is quite fast and accurate, but it does not support adding users or movies.

Field Summary

java.lang.String

FEATURE_FILE

Fields inherited from class netflix.recommender.AbstractRecommender

mh

Constructor Summary

IncRecommender(MemHelper mh)
Constructor.

IncRecommender(MemHelper mh, java.lang.String featureFile)
Constructor.

IncRecommender(java.lang.String mhFile)
Constructor.

Method Summary

double

recommend(int activeUser, int targetMovie, java.lang.String date) Predicts the rating for movie targetMovie by user activeUser.

Methods inherited from class netflix.recommender.AbstractRecommender

add, getUnratedMovies, getUnratedMovies, rankMovies, recommendFile, resort, testWithMemHelper, testWithMemHelper

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

FEATURE_FILE

public java.lang.String FEATURE_FILE

Constructor Detail

IncRecommender

public IncRecommender(java.lang.String mhFile)

Constructor. The MemHelper is not used, but is necessary for compatibility with AbstractRecommender.

Parameters:

mhFile - The file containing MemHelper object.

IncRecommender

public IncRecommender(MemHelper mh)

Constructor. The MemHelper is not used, but is necessary for compatibility with AbstractRecommender.

Parameters:

mh - The MemHelper object.

IncRecommender

public IncRecommender(MemHelper mh,
                      java.lang.String featureFile)

Constructor. The MemHelper is not used but is necessary for compatiblity with AbstractRecommender.

Parameters:

mh - The MemHelper object.

featureFile - Serialzed IncrementalSVDMovielens object.

Method Detail

recommend

public double recommend(int activeUser,
                        int targetMovie,
                        java.lang.String date)

Predicts the rating for movie targetMovie by user activeUser.

Specified by:

recommend in class AbstractRecommender

Parameters:

activeUser - User to predict rating for.

targetMovie - Movie to predict rating for.

date - The date the movie was rated (not used)

Returns:

The predicted rating.