Scholarlink

This work investigates the automatic evaluation of open-ended questions, the challenge of which is how to deal with natural language data. This opens up opportunities to explore Intended Learning Outcomes (ILO) more broadly. It is particularly beneficial for MOOCs, given the general lack of an instructor. We use Named Entity Recognition (NER), an NLP (natural language processing) approach, anchored in Bloom’s taxonomy with rubrics. A pilot test of a recently developed MOOC at LUT University in Finland was used as a case study. A total of 107 student responses were analyzed, with hit rates generally above 95%. Examples of using the NER system for student feedback in an adaptive learning environment are shown. Personalized pathways improve student learning and engagement, which also benefits MOOC completion.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Improving the Student Learning Process in MOOCs Through the Analysis of Open-Ended Question-Based Assessments Using Natural Language Processing

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Gustavo Almeida,
  • Johanna Naukkarinen,
  • Terhi Jantunen,
  • Soumya Datta,
  • Katja Kuparinen,
  • Esa Vakkilainen

摘要

This work investigates the automatic evaluation of open-ended questions, the challenge of which is how to deal with natural language data. This opens up opportunities to explore Intended Learning Outcomes (ILO) more broadly. It is particularly beneficial for MOOCs, given the general lack of an instructor. We use Named Entity Recognition (NER), an NLP (natural language processing) approach, anchored in Bloom’s taxonomy with rubrics. A pilot test of a recently developed MOOC at LUT University in Finland was used as a case study. A total of 107 student responses were analyzed, with hit rates generally above 95%. Examples of using the NER system for student feedback in an adaptive learning environment are shown. Personalized pathways improve student learning and engagement, which also benefits MOOC completion.