Automatic Prediction of the Relevance and Efficacy of College Students’ Online Learning Questions

• In Event: AECT - Automatic Prediction of the Relevance and Efficacy of College Students’ Online Learning Questions

Short Description

This study explores automatic classification of college students’ learning questions. 983 questions were collected from a Q & A platform implemented by four large programming course in a research university. Questions were manually classified into three categories: 1) learning-content-irrelevant, 2) effective learning-content-relevant, 3) ineffective learning-content-relevant. Four machine learning algorithms, including Naive Bayes Multinomial, Logistic Regression, Support Vector Machines, and Boosted Decision Tree, will be used for automatic classification and compared to the manual classification.

Authors

• Qiang Hao, University of Georgia

• Brad Barnes, University of Georgia

• Robert Maribe Branch, University of Georgia