and

1Prediction of Plastic Preform Temperature Profile and modeling perspectives The introduction of polyethylene terephthalate (PET) makes it possible to produce safe and functional plastic containers at a reasonable cost. The production of a PET container usually involves two stages. First, amorphous PET is injection molded into a preform. Then, the preform is reheated and blown into a container. Material distribution of the container, an important factor that determines the shape and functionality, can be controlled in the reheat and blow stage by adjusting heating parameters. Accurate prediction of PET preform temperature profile during reheat will facilitate understanding of the process to achieve optimal control. In this paper we investigate the various data mining paradigms that can create models of the domain. We also demonstrate how these models can be compared so that we can choose the best model available. The results show that the neuro-fuzzy approach is more efficient and effective, when compared with traditional modeling approaches