Using Support Vector Machines to Classify Student Attentiveness for the Development of Personalized Learning Systems

There have been many studies in which researchers have attempted to classify student attentiveness. Many of these approaches depended on a qualitative analysis and lacked any quantitative analysis. Therefore, this work is focused on bridging the gap between qualitative and quantitative approaches to classify student attentiveness. Thus, this research applies machine learning algorithms (K-means and SVM) to automatically classify students as attentive or inattentive using data from a consumer RGB-D sensor. Results of this research can be used to improve teaching strategies for instructors at all levels and can aid instructors in implementing personalized learning systems, which is a National Academy of Engineering Grand Challenge. This research applies machine learning algorithms to an educational setting. Data from these algorithms can be used by instructors to provide valuable feedback on the effectiveness of their instructional strategies and pedagogies. Instructors can use this feedback to improve their instructional strategies, and students will benefit by achieving improved learning and subject mastery. Ultimately, this will result in the students' increased ability to do work in their respective areas. Broadly, this work can help advance efforts in many areas of education and instruction. It is expected that improving instructional strategies and implementing personalized learning will help create more competent, capable, and prepared persons available for the future workforce.