Hybrid CAT Using Bayes Classification and Two-Parameter Model
Much research and implementation has been done in the field of adaptive learning, while many such platforms exist almost none of them have tackled the problem of maintainability of such high demand systems. This paper proposes a new system using naive Bayes classifier and two-parameter model of IRT to develop a low cost, easy to maintain, self-evolving test platform. The proposed model harnesses the knowledge of the community while implementing powerful test theory. The paper discusses in detail the major modules of the system along with the related theory. The proposed model incorporates machine learning and IRT to provide a state of the art system while still being a community powered platform. The scope of the proposed model is visited. This paper provides a direction and precedent for the development of a new breed of low maintenance high capability test platforms.
KeywordsItem response model Naive Bayes model CAT (Computer adaptive Test) Two-parameter model Recommendation system
- 2.Hambleton, R.K., Swaminathan, H.: IRT Principles and Applications. Kluwe-Nijhoff, Boston (1985)Google Scholar
- 3.Harris, D.: Comparison of 1-, 2- and 3- Parameter IRT Models (1984). https://doi.org/10.1111/j.1745-3992.1989.tb00313.x
- 5.Ng, A.Y., Jordan, M.I.: On Discriminative vs Generative Classifiers: A Comparison of Logistic Regression and Naïve Bayes (2002)Google Scholar
- 6.McCallum, A., Nigam, K.: A Comparison of Event Models for Naive Bayes Text Classification (1998)Google Scholar