Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood
Paaßen B, Göpfert C, Pinkwart N (2022)
In: Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022). Cristea AI, Brown C, Mitrovic T, Bosch N (Eds); 555–559.
Konferenzbeitrag | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Herausgeber*in
Cristea, Alexandra I.;
Brown, Chris;
Mitrovic, Tanja;
Bosch, Nigel
Einrichtung
Abstract / Bemerkung
Item response theory models the probability of correct student responses based on two interacting parameters: student ability and item difficulty. Whenever we estimate student ability, students have a legitimate interest in knowing how certain the estimate is. Confidence intervals are a natural measure of uncertainty. Unfortunately, computing confidence intervals can be computationally demanding. In this paper, we show that confidence intervals can be expressed as the solution to a feature relevance optimization problem. We use this insight to develop a novel solver for confidence intervals and thus achieve speedups by 4-50x while retaining near-indistinguishable results to the state-of-the-art approach.
Stichworte
item response theory;
confidence intervals;
relevance intervals;
approximation
Erscheinungsjahr
2022
Titel des Konferenzbandes
Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022)
Seite(n)
555–559
Konferenz
15th International Conference on Educational Data Mining (EDM 2022)
Konferenzort
Durham, UK
Konferenzdatum
2022-07-24 – 2022-07-27
Page URI
https://pub.uni-bielefeld.de/record/2979000
Zitieren
Paaßen B, Göpfert C, Pinkwart N. Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood. In: Cristea AI, Brown C, Mitrovic T, Bosch N, eds. Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022). 2022: 555–559.
Paaßen, B., Göpfert, C., & Pinkwart, N. (2022). Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood. In A. I. Cristea, C. Brown, T. Mitrovic, & N. Bosch (Eds.), Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022) (p. 555–559). https://doi.org/10.5281/zenodo.6852950
Paaßen, Benjamin, Göpfert, Christina, and Pinkwart, Niels. 2022. “Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood”. In Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022), ed. Alexandra I. Cristea, Chris Brown, Tanja Mitrovic, and Nigel Bosch, 555–559.
Paaßen, B., Göpfert, C., and Pinkwart, N. (2022). “Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood” in Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022), Cristea, A. I., Brown, C., Mitrovic, T., and Bosch, N. eds. 555–559.
Paaßen, B., Göpfert, C., & Pinkwart, N., 2022. Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood. In A. I. Cristea, et al., eds. Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022). pp. 555–559.
B. Paaßen, C. Göpfert, and N. Pinkwart, “Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood”, Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022), A.I. Cristea, et al., eds., 2022, pp.555–559.
Paaßen, B., Göpfert, C., Pinkwart, N.: Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood. In: Cristea, A.I., Brown, C., Mitrovic, T., and Bosch, N. (eds.) Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022). p. 555–559. (2022).
Paaßen, Benjamin, Göpfert, Christina, and Pinkwart, Niels. “Faster Confidence Intervals for Item Response Theory via an Approximate Likelihood”. Proceedings of the 15th International Conference on Educational Data Mining (EDM 2022). Ed. Alexandra I. Cristea, Chris Brown, Tanja Mitrovic, and Nigel Bosch. 2022. 555–559.
Open Access
Software:
Google Scholar