A Study on the Possibility of Automatically Estimating the Confidence Value of Students’ Knowledge in Generated Conceptual Models

摘要

—We propose a new metric to automatically evaluatethe confidence that a student knows a certain conceptincluded in his or her conceptual model. The conceptualmodel is defined as a simplified representation of theconcepts and relationships among them that a student keepsin his or her mind about an area of knowledge. Each area ofknowledge comprises several topics and each topic severalconcepts. Each concept can be identified by a term that thestudents should use. A concept can belong to one topic or toseveral topics. Terms are automatically extracted from theanswers provided to an automatic and adaptive free-textscoring system using Machine Learning techniques. In fact,the conceptual model is fully generated from the answersprovided by the students to this system. In the paper, theautomatic procedure that makes it possible is reviewed indetail. Finally, concept maps are used to graphically displaythe conceptual model to teachers and students. In this way,they can instantly see which concepts have already beenassimilated and which ones should still be reviewed.