Improving word-recognizers using an interactive lexicon with active and passive words

Abstract

The words a user is likely to write comprise the user's active vocabulary. This vocabulary is considerably smaller than the passive vocabulary of words a user reads. We explore an interactive adaptive lexicon method that separates a large lexicon into active and passive sets, and gradually expands and adapts the active set to reflect the user's active vocabulary. The adaptation is achieved through lightweight interaction as a by product of actual use. The effectiveness of the technique is demonstrated through a computational experiment and a user study. Copyright 2008 ACM.