Conformal sets in neural network regression

0384878 - ÚI 2013 RIV SK eng C - Konferenční příspěvek (zahraniční konf.)
Demut, R. - Holeňa, Martin
Conformal sets in neural network regression.
Information Technologies - Applications and Theory. Seňa: PONT s.r.o., 2012 - (Horváth, T.), s. 17-24. ISBN 978-80-971144-0-4.
[ITAT 2012. Conference on Theory and Practice of Information Technologies. Ždiar (SK), 17.09.2012-21.09.2012]
Grant CEP: GA ČR GA201/08/0802
Grant ostatní: GA CTU(CZ) SGS12/157/OHK4/2T/14
Institucionální podpora: RVO:67985807
Klíčová slova: nonlinear regression * artificial neural networks * confidence intervals * transductive inference * conformal sets
Kód oboru RIV: IN - Informatika

This paper is concerned with predictive regions in regression models, especially neural networks. We use the concept of conformal prediction (CP) to construct regions which satisfy given confidence level. Conformal prediction outputs regions, which are automatically valid, but their width and therefore usefulness depends on the used nonconformity measure. A nonconformity measure should tell us how different a given example is with respect to other examples. We de ne nonconformity measures based on some reliability estimates such as variance of a bagged model or local modeling of prediction error. We also present results of testing CP based on different nonconformity measures showing their usefulness and comparing them to traditional confidence intervals.
Trvalý link: http://hdl.handle.net/11104/0007330