KELDEC: A Recommendation System for Extending Classroom Learning with Visual Environmental Cues

Abstract

We develop an innovative personalized recommendation system called KELDEC that links the notes that students take in class with their outdoor experiences captured with camera, to suggest websites that extend their knowledge. Despite the plethora of educational recommendation systems, there is a dearth of effective tools that make evident the practical application of theory in the real world. KELDEC extracts the core learning points from class notes and distinctive labels that describe objects in a picture. It then mines the web to first extract the technical context of the picture, and subsequently culls out websites that establish linkages between notes and the picture. Response to user surveys garnered from students studying Software Engineering in the undergraduate Computer Engineering course reveal that they gain new and practical extension of classroom knowledge.