1 unstable release
0.1.0 | Jul 21, 2021 |
---|
#6 in #lsh
7KB
57 lines
hamming-bow
Produces binary term frequency bit arrays for hamming-space bag of words
How it works
This works by using hamming-dict
to create codewords in the hamming space that are as maximally spaced out as possible.
For each input key, its nearest neighbor is found in the dictionary and the corresponding bit is set in the bag. If the number of bits set in the bag becomes sufficiently large, the threshold number of word occurences required to set a bit will increase to balance the hash.
Dependencies
~340–580KB
~11K SLoC