Abstract: Emails, SMS, Chat transcripts and Social Network Messages play a vital role in present day’s communication. These digitally communicated data sets are different from well typed English documents like books, news papers …Etc in 2 ways. One is the use of acronyms instead of typing full form of the term and another one is the existence of misspelled words. Because of these reasons, information retrieval methods used for analyzing well typed English documents may not be well suitable to analyze them. In this paper we propose an information retrieval method to address the high recall requirements of retrieving information from digitally communicated data sets.