Adding LEPOR - A machine translation evaluation metric.

[ulhaqi12]

Hi,
Hope you are doing well. I have an idea of a potential contribution to the NLTK translation module.
Currently, in nltk.translate, we have bleu score, meteor score, and other evaluation metrics but there is a new method LEPOR that has better results than other evaluation metrics available in NLTK. I tried this when I was working on a machine translation project recently.
Lepor was not available in Python. After understanding the algorithm from the original paper, I have implemented this metric in Python and it is available here. I was thinking if can integrate this in NLTK as well so people can easily use it. In the future, the different variations of this metric will also be implemented.

Looking forward to your response.
Thank you

BR,
Ikram

[ulhaqi12]

Hey @stevenbird and @tomaarsen,

It would be great if I can get a green signal to generate PR.
Thank you

BR,
Ikram

[alvations]

@ulhaqi12 I'll suggest that you create the PR and I'll be glad to help you as a reviewer since I've interest in having good metrics added to the nltk.translate module.

[ulhaqi12]

Sure. I will do it. Thank you.