Collocation extraction toolin c#

Need bigram and trigram collocation extraction tool for English text written in c#. Can be standalone or as library - will eventually become part of a larger project. Collocation extraction should be based on the attached PDF document, also located here: [login to view URL] and similar to the NLTK software here: [login to view URL] Will require writing text parser and statistics extraction. Must be capable of analyzing large corpus of documents - tens of thousands of documents each with tens of thousands of words.