Doctor of Philosophy (PhD)
This dissertation details an intelligent tutoring system which aims to raise student trait abilities across concepts and levels of Blooms taxonomy within an educational program. The system utilizes Item Response Theory, a probabilistic theory of assessment which accounts for item discrimination, difficulty, and probability of guessing in its assessment of student trait ability. A modification to Item Response Theory which accounts for question dependencies is used to assess the probability that a student will answer a given question correctly given prior responses on the dependency questions. It also draws from theories of memory and forgetting present in the ACT-R model for intelligent tutoring systems. In particular, it models probability that a student will be able to recall the answer to a question outright, as well as the re-activation of memories of questions upon re-exposure. This model of memory is combined with the modified Item Response Theory to yield a more complete predictive model. Novel in this work is the fusion of Bloom’s taxonomy and Item Response Theory under the same system, graph-based model of dependency relationships among questions, mod- ified models of memory and forgetting, and fine-grained characterization-based models of student ability. In conjunction, these lend to efforts to schedule problems for the particular student which have the highest payoff.
Document Availability at the Time of Submission
Student has submitted appropriate documentation to restrict access to LSU for 365 days after which the document will be released for worldwide access.
Castleberry II, Dennis Gordon, "A Graph-Based Taxonomic Intelligent Tutoring System Utilizing
Bloom’S Taxonomy And Item-Response Theoretic Assessment" (2017). LSU Doctoral Dissertations. 4392.
Brandt, Steven R.