A three-step procedure for testing and validating the model is proposed:
firstly, the dataset is analyzed using projection methods as Principal Component Analysis (PCA)
and Cooperative Maximum-Likelihood Hebbian Learning (CMLHL) [1] to extract the dataset structure
and the key relations between variables and to know if the data set is sufficiently informative.
Finally, a model is produced at the modelling stage to estimate production time errors by modeling techniques.