dm.cs.tu-dortmund.de/mlbits/text-mining-search-retrieval/
Exact Text Search and Ranked Retrieval – Lecture Notes
mouse
rise
of
second
that
the
to
together
wealthy
wise
worm
Doc 1
1
1
1
1
1
1
Doc 2
1
1
1
1
1
1
1
1
Doc 3
1
1
1
1
1
1
Doc 4
1
1
1
1
1
1
1
1
1
1
1
intersection of bird AND early = 0011 AND 1010 = 0010
matching [...] store the quantity/positions along with the document:
bird
→
Doc 1 ×1, Doc 2 ×1
early
→
Doc 2 ×1, Doc 4 ×2
early
→
Doc 2 @ 4, Doc 4 @ 1,5
Ranked Retrieval – Approximative Matching
Exact matches (boolean [...] store the binary matrix for large document collections.
Solution: We store only the 1’s, for each term , sorted.
bird
→
Doc 1 , Doc 2
early
→
Doc 2 , Doc 4
Optimizations:
sort lists, and use a merge operation …