File Exchange

image thumbnail

Word Error Rate

version (8.67 KB) by Eduard Polityko
The function calculates WER between word sequences H (hypothesis) and R (reference).


Updated 07 Mar 2016

View License

Word error rate (WER) is a measure (metric) of the performance of an automatic speech recognition, machine translation etc.
The function is intended for calculation of WER between word sequence H (hypothesis) and word sequence R (reference).
For calculation we use Levenshtein distance on word level. Levenshtein distance is a minimal quantity of insertions, deletions and substitutions of words for conversion of a hypothesis to a reference. WER=D(H,R)/N, where D(H,R) is a Levenshtein distance between H and R and N is the number of words in the reference R. H and R are cell arrays of words (for example after using TEXTSCAN) or cells with word sequences or strings. Types H and R may be different.

Comments and Ratings (0)

MATLAB Release Compatibility
Created with R2015a
Compatible with any release
Platform Compatibility
Windows macOS Linux

Discover Live Editor

Create scripts with code, output, and formatted text in a single executable document.

Learn About Live Editor