News
PRG was awarded the New and Regular Faculty Internal Award for a machine learning project!
Machine learning can effectively transform chemistry education by allowing for computer-assisted analyses of student conceptual understanding on open-response assessments. The use of frequent formative, open-response assessments can allow instructors to effectively monitor student learning, provide students with feedback, as well as gather information to inform necessary changes to their instruction. Though open-response assessments are superior for revealing student thinking, instructors often resort to the use of multiple-choice assessments because they are easy to grade, particularly when teaching large classes. Machine learning can help address this problem by automating the assessment of students’ responses.
In this project, in collaboration with Dr. Shan Suthaharan (UNCG, Department of Computer Science), we aim to develop a computer model that can enable automated analysis of students’ explanations of infrared spectroscopy (IR) concepts and the prediction of students’ success when interpreting IR spectra. A recent study analyzing student understanding of IR concepts characterized five mental models that the students expressed when reasoning about IR spectroscopy and chemical structure. We propose to extract the words and phrases (features) that contribute to the computational descriptors of these mental models by using natural language processing (NLP) techniques, quantitative and computational techniques, and machine learning techniques. We will start with generating training and test data sets and will then use various techniques to derive a suitable machine learning model for classifying students that have successful interpretations of the IR spectra from those that do not. The IR spectroscopy machine learning model will enable instructors to promptly measure and monitor student learning, and (if needed) promptly and appropriately modify their teaching to support student learning.
Grateful to reviewers for their time and expertise, and to UNCG’s Office of Research and Engagement for financial support!
December 1, 2021
