Research9h ago
Unsupervised ML Maps Keta Basin Electrofacies
Six wireline logs from 11,195 samples in Ghana's Keta Basin provided the data for this study. Researchers used K-means clustering to identify four distinct electrofacies based on clay and porosity levels. The workflow achieves a 0.50 silhouette coefficient. This demonstrates how unsupervised learning characterizes rock properties when physical core data is unavailable.