Supervised Machine Learning in Electrofacies Classification: A Rough Set Theory Approach

Abstract

Electrofacies were initially introduced for defining a set of recorded log responses in order to characterize a bed and permitted it to be distinguished from the other rock units as an improvement to the traditional use of well logs. Grouping a formation into electrofacies can be used in lithology prediction, reservoir characterization and discrimination. Usually Multivariate statistical analyses, such as principal component analysis ‘PCA’ and cluster analysis are used for this purpose. In this study Extra Tree Classifier (ETC) based feature selection method is used to select the important attributes and three distinctive electrofacies were extracted from the dendrogram plot using the selected attributes. Finally, we proposed a rough set theory (RST) based white box classification approach to extract the pattern of the electrofacies in the form of decision rules which will allow the geosciences researchers to correlate the electrofacieses with the lithofacies from the extracted rough set (RS) rules.

Machine Learning in Electrofacies Classification and Subsurface Lithology Interpretation: A Rough Set Theory Approach

“…Recently, rule-based whitebox classification modules such as RST has been used in several related areas for solving classification problem analysis [30,31]. Touhid at al.…”