Assessing the aggregation of parameterized imprecise classification
Isabela Neves Drummond, Joaquím Meléndez, Sandra Aparecida Sandri · Conference on Artificial Intelligence Research and Development · 2006
This work is based on classifiers that can yield possibilistic valuations as output, that may have been obtained from a labeled data set either directly as such, by possibilistic classifiers, or by transforming the output of probabilistic classifiers or else by adapting prototype-based classifiers in general. Imprecise classifications are elicited from the possibilistic valuations by varying a parameter that makes the overall classification become either more or less precise. We discuss some accu-racy measures to assess the quality of the parameterized imprecise classifications, thus allowing the user to choose the most suitable level of imprecision for a given application. Here we particularly address the issue of aggregating parameterized aggregation classifiers, and assessing their performance.