Pattern recognition with evidential knowledge
M.A. Vila, Miguel Delgado, Antonio Skármeta · International Journal of Intelligent Systems · 1999
Many real situations may be seen as classifications problems if the information used to relate items with classes is a probability measure on 𝒫(X×C) (set of all subsets of X×C). It is well known that a probability measure on 𝒫(X×C) induces two evidence measures on X×C. We will formulate a classification model where it is assumed classes are fuzzy sets and the available information is evidential. Two bad classification loss functions will be established for this purpose, by using Dempster's upper and lower integrals. ©1999 John Wiley & Sons, Inc.