Lingua-EN-Syllable (0.251)

Lingua::EN::Syllable estimates the number of syllables in English words. In a test, it got about 87.6% right.

Here are the first few failure cases of the algorithm:

abandoned: 4 != 3
abased: 3 != 2
abasement: 4 != 3
abashed: 3 != 2
abatement: 4 != 3
abeyance: 2 != 3
aborigine: 4 != 5
abridged: 3 != 2
absorbed: 3 != 2
abstruseness: 4 != 3

Here the first number is the count given by Lingua::EN::Syllable and the second number is the correct number. As you can see, although it works moderately well, there are some obvious improvements which could be made, like not counting -ed at the end of words as a syllable unless there is a preceding t or d. However, it's likely that no purely algorithmic method could correctly count the syllables in words like "aborigine", so this is a fair stab. However, unfortunately, it doesn't have any way to return what it thinks the actual syllables are.

Ben Bullock - 2010-06-26T18:53:59