Familiarity breeds understanding Recommending explanatory analogies to learners

Explanatory analogies help learners in learning complex target concepts by mapping them onto more familiar source concept(s). Typically, multiple analogies are available for target concepts, though a limited number of them need to be recommended to the learner to reduce cognitive load during learning. While recommending explanatory analogies, it is essential that the learner be familiar with the source concept to facilitate learning, but the interest of the learner in a particular source concept may also play a role in making learning effective. In this paper, we present a survey with 94 participants to observe the role of familiarity and interest in choosing source concepts from multiple explanatory analogies. Furthermore, we present an approach to automatically identify source concept(s) from an analogy webpage. We then present an approach to recommend an explanatory analogy from a set of available analogies based on a familiarity measure derived for the source concept. We demonstrate this approach using a dataset of analogy webpages manually compiled from the internet.