Mining Open Answers in Questionnaire Data

and so aren’t satisfactorily reliable. Moreover, because the Web makes it so easy to take surveys and solicit comments, companies are finding themselves inundated with data from questionnaires and other sources. Handling it all manually would be not only cumbersome but also costly. Thus, devising a computer system that can automatically mine useful information from open answers has become an important issue. In general, processing answers in natural language is difficult because of the enormous variation in linguistic expression. A more realistic approach, and one that we believe can yield rather useful results, is to segment open answers into words and conduct an analysis at the word and phrase levels. We have developed a survey analysis system that works on these principles.1 The system, which we call SA in this article, is available under the name SurveyAnalyzer in Japan, where it is a trademark of NEC Corporation in Japan (see http://www1.ias.biglobe.ne.jp/ ipqos/solution/netvoyant/survey_analyzer.html). SA mines open answers through two statistical learning techniques: rule learning (which we call rule analysis) and correspondence analysis.