Automatic Classification of Harris Survey Questions: An Experiment in the Organization of Information

Abstract

This article describes an ongoing project in the automatic classification of Louis Harris survey questions. The purpose of the project is to explore the feasibility of automatically organizing the questions in a form that can be usefully presented to researchers and to investigate the problems associated with automatic classification of natural language text. In general, the experiment reported here supports the belief that an important application of automatic classification is to transform an intractable mass of data into a structure suitable for further human processing.