ABSTRACT
We very briefly discussed the scientific theory-driven modelling of materials systems in Chapter 2. In this chapter we begin to discuss the use of experimental or industrial data in materials modelling. Though the discussion is limited to statistical methods, still not in the domain of computational intelligence (CI), insight into this type of modelling is absolutely necessary to finally understand the CI tools and their applications. In the field of materials modelling, especially in the case of modelling complex metallic systems, statistical modelling concepts, particularly regression analysis, have been used in developing predictive models for materials properties, transformation temperatures and also process variables for the past eight decades, if not longer. After the introduction of artificial neural networks in the field of materials modelling, statistical regression analysis took a back seat in the development of new models. However, the existing models have huge practical relevance even in today’s world of metallurgy. But in this era of highcapacity computers, the statistical concepts in materials engineering have become more relevant than before in the form of data mining, taking the form and shape of ‘materials informatics’. This chapter therefore begins with a brief description of statistical linear regression techniques and the past applications of the techniques in the metallurgical field. The concept of data mining is then introduced, followed by a few accounts of its recent applications in the materials field.
