ANALYSIS OF FREE ANSWER IN QUESTIONNAIRE SURVEY AND EVALUATION OF LIVING ENVIRONMENT BY TEXT MINING
Yuji Kobayashi, Mitsunobu Terada, Seiji Sato · Journal of Architecture and Planning (Transactions of AIJ) · 2012
In this paper, the contents of the free answer in the questionnaire survey were analyzed by Text Mining, the residents's evaluation of living environment and the characteristics were grasped by the Correspondence Analysis and Cluster Analysis. And problems and possibility of application were considered in the questionnaire survey. Firstly, the free answers were sorted by seven categories, which are“Negative Evaluation”,“Demand”,“Negative Evaluation & Demand”and so on. And, the appearance frequency of words in the free answer were grasped. Next, using“Negative Evaluation”,“Demand”and“Negative Evaluation & Demand”on each category, the relationship with areas and appearance words were grasped by Correspondence Analysis, and the appearance patterns of the word were classified by Cluster Analysis. As a result, the residents's problems and demands in residence area were grasped, and a problem was arranged when a free answer will be analyzed.