Possibilities of "data mining" of production control in metallurgy for prediction of strength, ductility and toughness of metal products

Authors:

Abstract
The possibilities of a retrospective analysis of the production control database for predicting the level of strength, ductility and toughness of serial metal products have been evaluated. The boundaries of the effective application of regression, cognitive graphics techniques, parametric and non-parametric statistics (taking into account the type of distribution of values of control parameters) were revealed, which are necessary to search for significant relationships in the technology parameters - properties system and areas with a dominant type of dependence.