An Automatic Comparison between Knowledge Diagnostic Techniques

Previous works have pointed out the crucial need for comparison between knowledge diagnostic tools in the field of Intelligent Tutoring Systems (ITS). In this paper, we present an approach to compare knowledge diagnostics. We illustrate our proposition by applying three criteria of comparison for various diagnostic tools in geometry. K eywords: knowledge diagnostic, student modeling, comparison In the field of Intelligent Tutoring Systems (ITS), knowledge diagnostic use data collected from an ITS during interactions with the learner in order to infer the skills mastered or not by the student. Student modeling is a complex task: skills refer to a particular domain (mathematics one ITS and some specific tasks. Thus a comparison between two knowledge diagnostic processes is very difficult. Our work aims at assisting a user who wants to compare existing knowledge diagnostic tools in a particular domain, like benchmarks in computer sciences. Authors such as [1] have pointed out the crucial need for comparison as a way to improve knowledge diagnostic evaluation. We plan to give the same activity traces (i.e. record of all interactions of the student with an ITS in a particular domain) as input to various diagnostic tools, and to evaluate in an automatic way a set of criteria on their outputs, in order to assist the user in his comparison.