Research5h ago
Unsupervised ML Maps Keta Basin Electrofacies
Researchers analyzed 11,195 wireline log samples from Ghana's Keta Basin using K-means clustering. The workflow identified four distinct electrofacies with a 0.50 average silhouette coefficient. This unsupervised approach characterizes porosity and clay content without relying on scarce core data. It provides a scalable template for geological mapping in data-poor offshore environments.