The initial analysis of failures emerging in production process for further data mining analysis

Abstract

The aim of this paper is to examine possibilities for the initial data analyses of the failure data from industrial production process. To perform the initial data analysis of the data from production process we have used graphical statistical method and also data mining methods like drill-down analysis and cluster analysis. Before applying mentioned techniques and methods it was necessary to know the principle of the industrial production process itself and also to be aware of the failure data structure. This initial data analysis is vital to be able to review the knowledge potential of given data. Based on this, we are able to point out interesting issues, that can be further solved with KDD (knowledge discovery from databases) techniques.