Bootstrapping Accurate Learner Models from Electronic Portfolios

This paper presents a way to use electronic portfolios (e-portfolios) to initialize learner models for adaptive learning environments. E-portfolios become sources of evidence for claims about prior conceptual knowledge or skills. We developed the EP-LM system to test whether accurate learner models can be built from e-portfolios and to determine whether this approach can be beneficial for reflective learning. EP-LM supports e-portfolio browsing, claiming skill levels on certain concepts, and evidencing the claims using artifacts included in the e-portfolio. Experimental results show that accurate learner models can be created through this approach and that reflective and personalized learning can result.