As future work, it could be interesting to experiment the proposed approach with char n-grams instead
of word n-grams. This could help FASTDOCODE to include an intrinsic evaluation of a given document,
or help the algorithm to detect plagiarized passages with high obfuscation levels.