An incremental classification method Of questionnaire data using self-regulated judgment parameters
Yuki Mitsui, Kaoru Iida, Masanori Akiyoshi, Norihisa Komoda · 2010
This paper addresses a method to classify users' opinions into categories to analyze opinions from large amount of answers in open-ended questionnaires correctly. Our previous proposed system uses category classification samples as category-based dictionary, which has performance deterioration in case of a few samples, that is, “cold start problem”. This paper introduces a new incremental classification method with automatic updating for category classification samples by using self-regulating threshold values of judgment. We also discuss applied results of our proposed method to questionnaires about university lecture.