The Poisson Collocation Measure and its Applications

In this paper we introduce a measure for calculating statistically significant collocation sets that is related to the Poisson distribution. We show that results calculated using this measure are comparable to well-known measures like the log-likelihood measure. Additionally, we discuss asymptotic behaviour and additivity as general properties of the measure which may be applied to the analysis of multi-word collocations as well and can be used for defining a relative collocation measure as well. Finally, we give a brief overview of various possibilities of postprocessing collocation sets.