Research Publications

This paper explores the use of machine learning techniques to restore punctuation and case in English text, as part of which it investigates the co-dependence of case information and punctuation. We achieve an overall F-score of .619 for the task using a variety of lexical and contextual features, and iterative retagging.