Is hLepor available for all languages?

[nicolabertoldi]

Apart from the ngram-based version (nLEPOR_baseline), hLepor exploits language dependent morphological info.
In the paper, it is reported that, the morpho anlyzers are available for a few languages: cs,de,en, es,fr,ru.

What happens with the rest of the languages?

[poethan]

as long as the free part of speech and /or parsing tools are available, hLEPOR / LEPOR can be calculated on the word-surface level and part of speech sequence level then combine the two scores together. This was designed to address the synonym situation (e.g., some similarity between 'work' and 'works'), and sentence syntax structure similarity (eg.. the constituency parsed chunk similarity). But, without any morphology information, hLEPOR / LEPOR / ngram LEPOR can also be directly used on just word sequences for exact word match similarity, as long as there is system output and references.