Hi,

Is it a good approach to use Language Models, particularly unigram language model as a text similarity measure? Is there any clear advantage of VSM (or any other model) over LM in case of measuring text similarities? If so what are they?

I highly appreciate your kind comments.

Manjula.