Classification of Worries and Consultations with School Refusal Students Using Machine Learning
Rikuto Ishikura, Masahiro Takeda, Shino Iwashita · 2020
This study proposes a text classification method for the automatic tagging of posts on the an Internet bulletin board system for school refusal students. Four tags are defined on the basis of the reasons for truancy as follows: Friend, Teacher, Family, and Study. A classifier is generated for each tag, and multiple tags are assigned to a sentence. The results of classifier verification showed that the accuracy of each classifier was more than 70%, and the highest was more than 85%. However, the precision was low for the classifiers other than Family. Among the solutions, such as reviewing the dataset and changing the vectorizer, we will choose the method that is appropriate for each classifier.