What happens if you treat ordinal ratings as interval data? Human evaluations in NLP are even more under-powered than you think
David M. Howcroft, Verena Rieser
2021Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing17 citationsDOIOpen Access PDF
Abstract
Previous work has shown that human evaluations in NLP are notoriously under-powered.
Topics & Concepts
Variance (accounting)Ordinal regressionOrdinal dataComputer scienceArtificial intelligenceInterval (graph theory)Natural language processingMachine learningOrdinal ScalePower (physics)StatisticsMathematicsPhysicsQuantum mechanicsAccountingBusinessCombinatoricsNatural Language Processing TechniquesTopic ModelingAdvanced Text Analysis Techniques