Investigating Classification for Natural Language Processing Tasks

In this volume, Ben Medlock investigates the application of machine learning classification techniques to four key natural language processing tasks: topic classification, spam filtering, anonymisation and hedge identification. Theoretical arguments are combined with practical experiments, and consideration is given to the impact of the presented research within the wider context of the field as a whole. Natural language processing (NLP) is a thriving discipline in both the research and commercial arenas. It is central to the next generation of computational systems and a key technology in advancing the ease with which humans and machines interact. The classification paradigm, drawn from the field of machine learning, is a generic framework within which a computational learner induces a functional mapping between a particular sample space and a set of designated target classes, or in simple terms, the automatic assignment of category labels to data. This is a wide-ranging exploration of applied classification techniques for NLP and an essential read for those with an interest in the application of computers to language processing.