ABSTRACT
The discussion of data-driven techniques of modelling started in Chapter 3, where the statistical tools are discussed in brief. In this chapter we describe two computational intelligence (CI) techniques: artificial neural networks and genetic programming for empirical modelling, which have drawn significant applications in the field of materials modelling. As discussed in Chapter 3, the paradigm of such empirical modelling has shifted from its conventional approach of merely developing a predictive model to knowledge discovery using the data mining concept. In this perspective the CIbased tools have immense importance in modelling materials data, owing to their ability to capture the pattern of highly complex systems from the data. Techniques other than the two dealt with in this chapter, such as support vector machines, are also gradually gaining ground in the materials field. Another facet of data-driven modelling is rule-based modelling, where the rules are extracted from the data. This aspect is discussed in Chapter 5.
