/// @param n neighborhood size; capped at the number of users in the data model
/// @param minSimilarity minimal similarity required for neighbors
/// @throws IllegalArgumentException
/// if {@code n < 1}, or userSimilarity or dataModel are {@code null}
public NearestNUserNeighborhood(int n,
double minSimilarity,
IUserSimilarity userSimilarity,
IDataModel dataModel) :
this(n, minSimilarity, userSimilarity, dataModel, 1.0)
{
}
internal DataModelSimilaritiesIterator(IUserSimilarity otherSimilarity, long[] itemIDs)
{
this.otherSimilarity = otherSimilarity;
this.itemIDs = itemIDs;
i = 0;
itemID1 = itemIDs[0];
j = 1;
}
/// @param threshold
/// similarity threshold
/// @param userSimilarity
/// similarity metric
/// @param dataModel
/// data model
/// @param samplingRate
/// percentage of users to consider when building neighborhood -- decrease to trade quality for
/// performance
/// @throws IllegalArgumentException
/// if threshold or samplingRate is {@link Double#NaN}, or if samplingRate is not positive and less
/// than or equal to 1.0, or if userSimilarity or dataModel are {@code null}
public ThresholdUserNeighborhood(double threshold,
IUserSimilarity userSimilarity,
IDataModel dataModel,
double samplingRate)
: base(userSimilarity, dataModel, samplingRate) {
//Preconditions.checkArgument(!Double.isNaN(threshold), "threshold must not be NaN");
this.threshold = threshold;
}
/// @param threshold
/// similarity threshold
/// @param userSimilarity
/// similarity metric
/// @param dataModel
/// data model
/// @param samplingRate
/// percentage of users to consider when building neighborhood -- decrease to trade quality for
/// performance
/// @throws IllegalArgumentException
/// if threshold or samplingRate is {@link Double#NaN}, or if samplingRate is not positive and less
/// than or equal to 1.0, or if userSimilarity or dataModel are {@code null}
public ThresholdUserNeighborhood(double threshold,
IUserSimilarity userSimilarity,
IDataModel dataModel,
double samplingRate)
: base(userSimilarity, dataModel, samplingRate)
{
//Preconditions.checkArgument(!Double.isNaN(threshold), "threshold must not be NaN");
this.threshold = threshold;
}
public GenericUserSimilarity(IUserSimilarity otherSimilarity,
IDataModel dataModel,
int maxToKeep)
{
long[] userIDs = longIteratorToList(dataModel.GetUserIDs());
IEnumerator <UserUserSimilarity> it = new DataModelSimilaritiesIterator(otherSimilarity, userIDs);
var keptSimilarities = TopItems.GetTopUserUserSimilarities(maxToKeep, it);
initSimilarityMaps(keptSimilarities.GetEnumerator());
}
/// Creates this on top of the given {@link UserSimilarity}.
/// The cache size is capped by the given size.
public CachingUserSimilarity(IUserSimilarity similarity, int maxCacheSize)
{
//Preconditions.checkArgument(similarity != null, "similarity is null");
this.similarity = similarity;
this.similarityCache = new Cache <Tuple <long, long>, Double>(new SimilarityRetriever(similarity), maxCacheSize);
this.refreshHelper = new RefreshHelper(() => {
similarityCache.Clear();
});
refreshHelper.AddDependency(similarity);
}
public AbstractUserNeighborhood(IUserSimilarity userSimilarity, IDataModel dataModel, double samplingRate) {
//Preconditions.checkArgument(userSimilarity != null, "userSimilarity is null");
//Preconditions.checkArgument(dataModel != null, "dataModel is null");
//Preconditions.checkArgument(samplingRate > 0.0 && samplingRate <= 1.0, "samplingRate must be in (0,1]");
this.userSimilarity = userSimilarity;
this.dataModel = dataModel;
this.samplingRate = samplingRate;
this.refreshHelper = new RefreshHelper(null);
this.refreshHelper.AddDependency(this.dataModel);
this.refreshHelper.AddDependency(this.userSimilarity);
}
public override long[] GetUserNeighborhood(long userID)
{
IDataModel dataModel = getDataModel();
IUserSimilarity userSimilarityImpl = getUserSimilarity();
TopItems.IEstimator <long> estimator = new Estimator(userSimilarityImpl, userID, minSimilarity);
var userIDs = SamplinglongPrimitiveIterator.MaybeWrapIterator(dataModel.GetUserIDs(),
getSamplingRate());
return(TopItems.GetTopUsers(n, userIDs, null, estimator));
}
public AbstractUserNeighborhood(IUserSimilarity userSimilarity, IDataModel dataModel, double samplingRate)
{
//Preconditions.checkArgument(userSimilarity != null, "userSimilarity is null");
//Preconditions.checkArgument(dataModel != null, "dataModel is null");
//Preconditions.checkArgument(samplingRate > 0.0 && samplingRate <= 1.0, "samplingRate must be in (0,1]");
this.userSimilarity = userSimilarity;
this.dataModel = dataModel;
this.samplingRate = samplingRate;
this.refreshHelper = new RefreshHelper(null);
this.refreshHelper.AddDependency(this.dataModel);
this.refreshHelper.AddDependency(this.userSimilarity);
}
/// @param n neighborhood size; capped at the number of users in the data model
/// @param minSimilarity minimal similarity required for neighbors
/// @param samplingRate percentage of users to consider when building neighborhood -- decrease to trade quality for
/// performance
/// @throws IllegalArgumentException
/// if {@code n < 1} or samplingRate is NaN or not in (0,1], or userSimilarity or dataModel are
/// {@code null}
public NearestNUserNeighborhood(int n,
double minSimilarity,
IUserSimilarity userSimilarity,
IDataModel dataModel,
double samplingRate)
: base(userSimilarity, dataModel, samplingRate)
{
//Preconditions.checkArgument(n >= 1, "n must be at least 1");
int numUsers = dataModel.GetNumUsers();
this.n = n > numUsers ? numUsers : n;
this.minSimilarity = minSimilarity;
}
public GenericUserBasedRecommender(IDataModel dataModel,
IUserNeighborhood neighborhood,
IUserSimilarity similarity) : base(dataModel)
{
//Preconditions.checkArgument(neighborhood != null, "neighborhood is null");
this.neighborhood = neighborhood;
this.similarity = similarity;
this.refreshHelper = new RefreshHelper(() => {
capper = buildCapper();
});
refreshHelper.AddDependency(dataModel);
refreshHelper.AddDependency(similarity);
refreshHelper.AddDependency(neighborhood);
capper = buildCapper();
}
/// This computation is in a technical sense, wrong, since in the domain of "bool preference users" where
/// all preference values are 1, this method should only ever return 1.0 or NaN. This isn't terribly useful
/// however since it means results can't be ranked by preference value (all are 1). So instead this returns a
/// sum of similarities to any other user in the neighborhood who has also rated the item.
protected override float doEstimatePreference(long theUserID, long[] theNeighborhood, long itemID)
{
if (theNeighborhood.Length == 0)
{
return(float.NaN);
}
IDataModel dataModel = GetDataModel();
IUserSimilarity similarity = getSimilarity();
float totalSimilarity = 0.0f;
bool foundAPref = false;
foreach (long userID in theNeighborhood)
{
// See GenericItemBasedRecommender.doEstimatePreference() too
if (userID != theUserID && dataModel.GetPreferenceValue(userID, itemID) != null)
{
foundAPref = true;
totalSimilarity += (float)similarity.UserSimilarity(theUserID, userID);
}
}
return(foundAPref ? totalSimilarity : float.NaN);
}
public override long[] GetUserNeighborhood(long userID)
{
IDataModel dataModel = getDataModel();
FastIDSet neighborhood = new FastIDSet();
var usersIterable = SamplinglongPrimitiveIterator.MaybeWrapIterator(dataModel
.GetUserIDs(), getSamplingRate());
IUserSimilarity userSimilarityImpl = getUserSimilarity();
while (usersIterable.MoveNext())
{
long otherUserID = usersIterable.Current;
if (userID != otherUserID)
{
double theSimilarity = userSimilarityImpl.UserSimilarity(userID, otherUserID);
if (!Double.IsNaN(theSimilarity) && theSimilarity >= threshold)
{
neighborhood.Add(otherUserID);
}
}
}
return(neighborhood.ToArray());
}
/// @param threshold
/// similarity threshold
/// @param userSimilarity
/// similarity metric
/// @param dataModel
/// data model
/// @throws IllegalArgumentException
/// if threshold is {@link Double#NaN}, or if samplingRate is not positive and less than or equal
/// to 1.0, or if userSimilarity or dataModel are {@code null}
public ThresholdUserNeighborhood(double threshold, IUserSimilarity userSimilarity, IDataModel dataModel) :
this(threshold, userSimilarity, dataModel, 1.0) {
}
/// @param n neighborhood size; capped at the number of users in the data model
/// @throws IllegalArgumentException
/// if {@code n < 1}, or userSimilarity or dataModel are {@code null}
public NearestNUserNeighborhood(int n, IUserSimilarity userSimilarity, IDataModel dataModel) :
this(n, Double.NegativeInfinity, userSimilarity, dataModel, 1.0)
{
}
internal Estimator(IUserSimilarity userSimilarityImpl, long theUserID, double minSim)
{
this.userSimilarityImpl = userSimilarityImpl;
this.theUserID = theUserID;
this.minSim = minSim;
}
internal SimilarityRetriever(IUserSimilarity similarity)
{
this.similarity = similarity;
}
public GenericBooleanPrefUserBasedRecommender(IDataModel dataModel,
IUserNeighborhood neighborhood,
IUserSimilarity similarity)
: base(dataModel, neighborhood, similarity) {
}
/// Creates this on top of the given {@link UserSimilarity}.
/// The cache is sized according to properties of the given {@link DataModel}.
public CachingUserSimilarity(IUserSimilarity similarity, IDataModel dataModel) : this(similarity, dataModel.GetNumUsers())
{
}
public GenericBooleanPrefUserBasedRecommender(IDataModel dataModel,
IUserNeighborhood neighborhood,
IUserSimilarity similarity)
: base(dataModel, neighborhood, similarity)
{
}
/// @param threshold
/// similarity threshold
/// @param userSimilarity
/// similarity metric
/// @param dataModel
/// data model
/// @throws IllegalArgumentException
/// if threshold is {@link Double#NaN}, or if samplingRate is not positive and less than or equal
/// to 1.0, or if userSimilarity or dataModel are {@code null}
public ThresholdUserNeighborhood(double threshold, IUserSimilarity userSimilarity, IDataModel dataModel) :
this(threshold, userSimilarity, dataModel, 1.0)
{
}
public GenericUserSimilarity(IUserSimilarity otherSimilarity, IDataModel dataModel)
{
long[] userIDs = longIteratorToList(dataModel.GetUserIDs());
initSimilarityMaps(new DataModelSimilaritiesIterator(otherSimilarity, userIDs));
}
internal MostSimilarEstimator(long toUserID, IUserSimilarity similarity, IRescorer <Tuple <long, long> > rescorer)
{
this.toUserID = toUserID;
this.similarity = similarity;
this.rescorer = rescorer;
}
