%0 Generic %1 lan2013joint %A Lan, Andrew S. %A Studer, Christoph %A Waters, Andrew E. %A Baraniuk, Richard G. %D 2013 %K joint model topic toread %T Joint Topic Modeling and Factor Analysis of Textual Information and Graded Response Data %U http://arxiv.org/abs/1305.1956 %X Modern machine learning methods are critical to the development of large-scale personalized learning systems that cater directly to the needs of individual learners. The recently developed SPARse Factor Analysis (SPARFA) framework provides a new statistical model and algorithms for machine learning-based learning analytics, which estimate a learner's knowledge of the latent concepts underlying a domain, and content analytics, which estimate the relationships among a collection of questions and the latent concepts. SPARFA estimates these quantities given only the binary-valued graded responses to a collection of questions. In order to better interpret the estimated latent concepts, SPARFA relies on a post-processing step that utilizes user-defined tags (e.g., topics or keywords) available for each question. In this paper, we relax the need for user-defined tags by extending SPARFA to jointly process both graded learner responses and the text of each question and its associated answer(s) or other feedback. Our purely data-driven approach (i) enhances the interpretability of the estimated latent concepts without the need of explicitly generating a set of tags or performing a post-processing step, (ii) improves the prediction performance of SPARFA, and (iii) scales to large test/assessments where human annotation would prove burdensome. We demonstrate the efficacy of the proposed approach on two real educational datasets.

@misc{lan2013joint, abstract = {Modern machine learning methods are critical to the development of large-scale personalized learning systems that cater directly to the needs of individual learners. The recently developed SPARse Factor Analysis (SPARFA) framework provides a new statistical model and algorithms for machine learning-based learning analytics, which estimate a learner's knowledge of the latent concepts underlying a domain, and content analytics, which estimate the relationships among a collection of questions and the latent concepts. SPARFA estimates these quantities given only the binary-valued graded responses to a collection of questions. In order to better interpret the estimated latent concepts, SPARFA relies on a post-processing step that utilizes user-defined tags (e.g., topics or keywords) available for each question. In this paper, we relax the need for user-defined tags by extending SPARFA to jointly process both graded learner responses and the text of each question and its associated answer(s) or other feedback. Our purely data-driven approach (i) enhances the interpretability of the estimated latent concepts without the need of explicitly generating a set of tags or performing a post-processing step, (ii) improves the prediction performance of SPARFA, and (iii) scales to large test/assessments where human annotation would prove burdensome. We demonstrate the efficacy of the proposed approach on two real educational datasets.}, added-at = {2013-10-18T13:27:38.000+0200}, author = {Lan, Andrew S. and Studer, Christoph and Waters, Andrew E. and Baraniuk, Richard G.}, biburl = {https://www.bibsonomy.org/bibtex/22a8df43258181ed85e5d43b489fd45fb/hotho}, description = {Joint Topic Modeling and Factor Analysis of Textual Information and Graded Response Data}, interhash = {911707523671c994e5c3fe63c3df5c4a}, intrahash = {2a8df43258181ed85e5d43b489fd45fb}, keywords = {joint model topic toread}, note = {cite arxiv:1305.1956}, timestamp = {2013-10-18T13:27:38.000+0200}, title = {Joint Topic Modeling and Factor Analysis of Textual Information and Graded Response Data}, url = {http://arxiv.org/abs/1305.1956}, year = 2013 }