Analytical Solution for the Optimal Addition of an Item to a Composite of Scores for Maximum Reliability
Ferrer CA, Torres-Rodríguez I, Taboada-Crispi A, Nöth E (2019)
Publication Type: Conference contribution
Publication year: 2019
Journal
Publisher: Springer
Book Volume: 11896 LNCS
Pages Range: 408-416
Conference Proceedings Title: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Event location: Havana
ISBN: 9783030339036
DOI: 10.1007/978-3-030-33904-3_38
Abstract
This paper presents a derivation of the optimal weight to be assigned for an item so that it maximally increases the reliability of the aggregate. This aggregate is the best estimate of the underlying true repeating pattern. The approach differs from previous solutions in being analytical, based on the Signal to Noise Ratio (SNR) instead of the reliability itself, and the ability to visually inform the researcher about the relevance of the weighting strategy and the gains produced in the SNR. Optimal weighting of repetitive phenomena is a bonus not only in the behavioral sciences, but also in many engineering fields. Its uses may include the selection or discarding of raters, judges, repetitions, or epochs, depending on the field.
Authors with CRIS profile
Involved external institutions
How to cite
APA:
Ferrer, C.A., Torres-Rodríguez, I., Taboada-Crispi, A., & Nöth, E. (2019). Analytical Solution for the Optimal Addition of an Item to a Composite of Scores for Maximum Reliability. In Ingela Nyström, Yanio Hernández Heredia, Vladimir Milián Núñez (Eds.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 408-416). Havana, CU: Springer.
MLA:
Ferrer, Carlos A., et al. "Analytical Solution for the Optimal Addition of an Item to a Composite of Scores for Maximum Reliability." Proceedings of the 24th Iberoamerican Congress on Pattern Recognition, CIARP 2019, Havana Ed. Ingela Nyström, Yanio Hernández Heredia, Vladimir Milián Núñez, Springer, 2019. 408-416.
BibTeX: Download