Rcnk rc"ctc tahmin1cmcsi; veri len biT rderans rcnk ilc aym rcngi urctmck iyin hangi boyamladdclcrin hangi konsantrasyonlarma ihtiya~ duyuldugunu tahmin ctmcktir. iyi biT tahminlcmc yapabilmck dcncyimli boya uzmanlan i"in bile zordur. R~C1C tahmini; klasik rc\,etc ~lkarma sistcminc gore veya bilgisayarh rc,"c\c !,'lkam13 sistcminc gore yapllabilir. Klasik rc,"cic vlkanna sistcmindc biT boya uzmalll dcncyimlcrinc dayanarak vcrilcn renge yaklll biT r~ctc tuttunnaya (j:ah~lr, bilgisayarh rcvctc (j:lkarma sistcmlcri isc yaygm olarak bilincn dogrusal modele (Kubclka-Munk modcli) veya dogrusal olmayan (yapay sinir aglafl) modele dayanabilir. Bu makalede dogrusal olmayan sisteme dayanan yapay sinir aglarmm kullanl ml ile reo;ete O;lkanna sisteminin nasi! ~

Color recipe prediction was known to predict which colorants and their concentrations were required for a given reference color. Making a good color recipe prediction is even hard to expen colorists. Color recipe prediction can be made by classical and computering recipe prediction systems. The classical system is based on a colorist's prediction, while the latter is applied as linear model (Kubelka-Munk model) in common or nonlinear mode! (such as neural network model). In this paper, how color recipe prediction by neural network based on nonlinear model is working was reviewed.