pEmpirical data used from corpora enable a linguist to make the statements that are objective and based on natural language. Therefore, the roles of corpora are pivotal in many different felds of study where language is the central object. The most widely used methodfor distinguishing between random co-occurrences and true collocations is the application of association measures. Those measures compute an association scorefor each word pair, which can then be usedfor ranking. One of the best-known association measures used for collocational studies in corpus linguistics is MI score (Mutual Information) suggested by Church and Hanks (1990). The focus of this paper is to fnd the most common verb-noun collocationsfrom the Corpus of Contemporary American English with the help of association measure MI. Moreover, the results are compared to thosefound in the British National Corpusfor the sake of high data accuracy./p
Mirna Begagić (Wed,) studied this question.