R character hash function -
ok question related one: feature hashing in r text classification. what's reasonable way hash character vector
integer
in r? current code take lastly hex chars , convert then, doesn't seems extract features (i've tested in simple classification problem):
library(digest) bad_hash = function(word, n){ strtoi(substr(digest(word), 28, 32), 16l) %% (n + 1) } bad_hash('this sparta', 100) # 97
thanks!
r hash text-classification
No comments:
Post a Comment