The Implementation of Machine Learning in Lithofacies Classification using Multi Well Logs Data

Lithofacies classification is a process to identify rock lithology by indirect measurements. Usually, the processed manually an experienced geoscientist. This research presents automated lithofacies using machine learning method increase computational power in shortening process's time consumption. The support vector (SVM) algorithm has been applied successfully Damar field, Indonesia. input various well-log data sets, e.g., gamma-ray, density, resistivity, neutron porosity, and effective porosity. Machine can classify seven depositional environments, including channel, bar sand, beach carbonate, volcanic, shale. accuracy verification phase with trained class reached more than 90%, while validation beyond 65%. classified then be used as for describing lateral vertical distribution patterns.