Abstract
Hanadayori provides advice by considering the readers of e-mails written by students for tasks. The system conducts a morphological analysis of the text and checks it against a stylistic list, mistaken use list by task, and a list of necessary expressions by task. If the expressions needed for the task are missing, it indicates mistaken usage and provides suggestions. An operational experiment using student data indicated that stylistic errors and advice on necessary expressions were as anticipated. However, sometimes, findings for mistaken usage could not be displayed. Hanadayori should henceforth become more useful by adding to the mistaken use list.
