Sally generates LibSVM vector spaces like
1 1191618:0.973645 1243483:0.362496 1346494:0.995527
0 1243483:0.226674 1346494:0.997515 2362908:1 4604526:1
where feature indices are really (murmur3?) hashes of words.
Maybe we could do
Svm::Model#predict(1191618 => 0.973645, 1243483 => 0.362496 [etc]
Sally generates LibSVM vector spaces like
where feature indices are really (murmur3?) hashes of words.
Maybe we could do