A Model for Function Word Counts
Brian J. Bailey · Journal of the Royal Statistical Society Series C (Applied Statistics) · 1990
SUMMARY Researchers in linguistics have considered an author's use of function words, or context-free words, as providing one measure of his style. The number of occurrences of a single word, such as 'and', or an uncommon group of words, in a fixed sample of n words, can often be modelled by a binomial distribution. However, this model sometimes breaks down for more common groups of words, such as all the articles, a group of pronouns etc. For such cases, a more realistic model is presented and analysed, based on a simple Markov chain model of English usage.