inst/benchmarks/ngonly.r

In wrathematics/ngram: Fast n-Gram 'Tokenization'

library(tau)
library(ngram)

x <- ngram::rcorpus(100000)

tautime <- system.time(pt1 <- textcnt(x, n=3, split=" ", method="string"))[3]
ngtime <- system.time(pt2 <- ngram(x, n=3))[3]

cat("tau: ", tautime, "\n")
cat("ngram: ", ngtime, "\n")
cat("tau/ngram: ", tautime/ngtime, "\n")