Sparse Matrix (Compressed Sparse Row, CSR)
Feature matrices from Concept - Bag of Words are mostly zeros, since most documents contain only a small subset of the vocabulary (02 Pre-processing s.70). Instead of storing all zeros, represent each row as a list of pairs (column number, nonzero value).
| Dense row | Sparse pairs |
|---|---|
| 0 15 0 3 | (2,15) (4,3) |
| 12 0 0 0 | (1,12) |
| 8 0 5 2 | (1,8) (3,5) (4,2) |
(Columns numbered from 1, as on the slide.)
Related
Concept - Document-Term Matrix | Concept - Curse of Dimensionality | Concept - Dictionary Reduction