org.grouplens.lenskit.transform.normalize

Interface UserVectorNormalizer

All Known Implementing Classes:

AbstractUserVectorNormalizer, BaselineSubtractingUserVectorNormalizer, DefaultUserVectorNormalizer

@DefaultImplementation(value=DefaultUserVectorNormalizer.class)
public interface UserVectorNormalizer

Normalize a user's vector. This vector is typically a rating or purchase vector.

This interface is essentially a user-aware version of VectorNormalizer. The default implementation, DefaultUserVectorNormalizer, delegates to a VectorNormalizer. Implement this interface directly to create a normalizer that is aware of the fact that it is normalizing a user and e.g. uses user properties outside the vector to aid in the normalization. Otherwise, use a context-sensitive binding of VectorNormalizer to configure the user vector normalizer:

factory.in(UserVectorNormalizer.class) .bind(VectorNormalizer.class) .to(MeanVarianceNormalizer.class);

Since:

0.11

See Also:

VectorNormalizer

Method Summary

Modifier and Type	Method and Description
VectorTransformation	makeTransformation(long user, SparseVector vector) Make a vector transformation for a user.
MutableSparseVector	normalize(long user, SparseVector vector, MutableSparseVector target) Normalize a vector with respect to a user vector.

Method Detail

normalize

MutableSparseVector normalize(long user,
                              @Nonnull
                              SparseVector vector,
                              @Nullable
                              MutableSparseVector target)

Normalize a vector with respect to a user vector.

Parameters:

user - The user to normalize a vector for.

vector - The user's vector for reference.

target - The vector to normalize. If null, the user vector is normalized.

Returns:

The target vector, if specified. Otherwise, a fresh mutable vector containing a normalized copy of the user vector is returned.

makeTransformation

VectorTransformation makeTransformation(long user,
                                        SparseVector vector)

Make a vector transformation for a user. The resulting transformation will be applied to user vectors to normalize and denormalize them.

Parameters:

user - The user ID to normalize for.

vector - The user's vector to use as the reference vector.

Returns:

The vector transformaition normalizing for this user.