|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |
public interface DistributedVectorSimilarity
a measure for the pairwise similarity of two rows of a matrix that is suitable for computing that similarity in a distributed way works in 2 steps: - at first weight() is called for each of the row vectors - later similarity is called with the previously computed weights as parameters
Method Summary | |
---|---|
double |
similarity(int rowA,
int rowB,
java.lang.Iterable<Cooccurrence> cooccurrences,
double weightOfVectorA,
double weightOfVectorB,
int numberOfColumns)
compute the similarity of a pair of row vectors |
double |
weight(Vector v)
compute the weight (e.g. |
Method Detail |
---|
double weight(Vector v)
double similarity(int rowA, int rowB, java.lang.Iterable<Cooccurrence> cooccurrences, double weightOfVectorA, double weightOfVectorB, int numberOfColumns)
rowA
- offset of the first rowrowB
- offset of the second rowcooccurrences
- all column entries where both vectors have a nonZero entryweightOfVectorA
- the result of weight(Vector)
for the first row vectorweightOfVectorB
- the result of weight(Vector)
for the first row vector
|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |