Applications of Feature Selection and Regression Techniques in Materials Design
Feature selection is considered as an important preprocessing step to data mining and soft computing, whereas regression is a collection of methods to optimally assess the signal from a noisy output. Both seek to arrive at the dependence and relation between different attributes and a target material property. In the present chapter a flock of regression and feature selection techniques are discussed, and the kind of results that can be obtained with each of them has been illustrated with the help of a dataset on steel. The different methods are capable of abstracting data in different forms, thus revealing hidden knowledge from different perspectives. Choosing the most appropriate method depends on the application at hand and the kind of objective that one is looking for.