<![CDATA[This research was conducted in the "VISA" field of the Balikpapan Formation, located in the Kutai Basin, one of Indonesia's largest hydrocarbon basins. The lithology of this formation is primarily Sandstone and shale, which are significant for hydrocarbon exploration and production. Determining the initial lithology is an essential step for understanding the characteristics of the well data during processing. Consequently, the Support Vector Machine (SVM) algorithm was implemented in this study to predict lithology using well data. This investigation employs data from four wells: VISA-9, VISA-13, VISA-36, and VISA-39. The prediction results are subsequently visualized as well as logs and lithology distribution histograms to make the results easier to interpret based on three interpreted lithology categories: Sandstone, shale, and Coal. Performance evaluations indicate that limitations remain in the SVM classification. The error range obtained in Experiments 1 and 2 was 11–22% compared to the actual lithology. However, Experiment 3 demonstrated substantial improvement by utilizing three training datasets, which reduced the error rate to 5% (a 7% improvement from previous experiments). Overall, the SVM method can effectively classify rock lithology; however, the model still requires optimization to minimize residual errors during the prediction process. Ultimately, this investigation demonstrates that SVM can be successfully applied to predict lithology using well log parameters.]]>
Copyrights © 2026
