An evaluation of different statistical techniques of collocation extraction using a probability measure to word combinations
Raj Kishor Bisht, Hoshiyar S. Dhami, Neeraj Tiwari · Journal of Quantitative Linguistics · 2006
A collocation is a recurrent combination of words which appears more often than by chance. Collocations play an important role in many natural language processing tasks such as computational lexicography, word sense disambiguation, machine translation, information retrieval, etc. In this paper an attempt has been made to evaluate different collocation extraction techniques for a small corpus. For this purpose, some word pairs have been taken from the corpus and results of each technique have been shown and discussed. Finally, a probability measure has been suggested to filter out some free word combinations before applying these tests, which in turn improves the results obtained from the algorithms.