抄録
This paper addresses a classification method of open-ended questionnaire using category-based dictionary from categorized samples. Different from generally-used document classification methods, our proposed method here introduces a new category-based dictionary which is generated from a small set of categorized samples, because the purpose of analysis is to have meaningful results along with a user's viewpoints in a sense. This category-based dictionary is used to judge a category as to a questionnaire with tf-idf (term frequency inverted document frequency) and co_tf-idf (co-occurrence tf-idf). We show the effectiveness in applying our proposed method to questionnaires about an university lecture.
