We previously reported a class questionnaire response analysis system (class QRAS) that can classify and retrieve student questionnaire responses collected from students during a class on the basis of semantic-distributed representations in real time. In this study, we added a function for classifying questionnaire responses into positive, negative, and neutral categories using the Japanese Sentiment Polarity Dictionary to the class QRAS. Using this function, it is possible to classify questionnaire responses, for example, into categories showing reasons why some are interested in the contents of a class and some are not. This enables the class QRAS to provide more advanced class support to the user.