Improving Upon Semantic Classification of Spoken Diary Entries Using Pragmatic Context Information

摘要

In this paper we will discuss Natural Language Processing (NLP) techniques used to analyze spoken diary entries and automatically classify the locations and activities into semantic categories. This research combines work done in utterance semantics classification using a Bayesian statistical model of the words with contextual information present within the diary to disambiguate classification of these utterances and achieve higher classification accuracy than a model that only relies on word analysis.