低资源和不平衡数据集下的强大教育对话行为分类器（英文版）

英文标题：Robust Educational Dialogue Act Classifiers with Low-Resource and Imbalanced Datasets中文摘要：本文介绍了如何通过优化样本分布的方式，提高智能辅导系统中对话行为（DA）分类器的准确性和鲁棒性，通过大规模实验进行验证，取得了良好的结果。英文摘要：Dialogue acts (DAs) can represent conversational actions of tutors orstudents that take place during tutoring dialogues. Automating theidentification of DAs in tutoring dialogues is significant to the design ofdialogue-based intelligent tutoring systems. Many prior studies employ machinelearning models to classify DAs in tutorin