Word Error Rate

link: range: [0, 1] (lesser the better)

summary

• Used to evaluate how good your model solves machine translation or transcription problems It’s derived from the Levenshtein distance

$WER=\frac{S+D+I}{N}=\frac{S+D+I}{S+D+C}$

• S is the number of substitutions,
• D is the number of deletions,
• I is the number of insertions,
• C is the number of correct words,
• N is the number of words in the reference (N=S+D+C)

pros

Cons