Well-Grouped Validation and Explainable XGBoost Lithology Classification under Missing Well-Log Conditions
DOI:
https://doi.org/10.54691/jn8f5g18Keywords:
Lithology Identification; Well-level Cross-validation; Logging Curves; XGBoost; Class Imbalance; Out-of-fold Prediction; SHAP.Abstract
To address challenges in actual logging interpretation—such as imbalanced lithology types, incomplete logging curves, and the difficulty of reliably evaluating cross-well generalization—this study performs intelligent lithology identification using the publicly available FORCE 2020 logging dataset. After preprocessing, the dataset comprises 98 wells, 1,170,511 depth samples, and 12 lithology types. Five conventional logging curves—natural gamma ray, bulk density, neutron porosity, acoustic transit time (DTC), and deep laterolog resistivity (RDEP)—are selected as input features. A 5-fold GroupKFold cross-validation ensures complete separation of training and test wells. XGBoost, LightGBM, and random forest classifiers are compared, together with various missing-value handling and class-weighting strategies. On this basis, model robustness to incomplete logging data is evaluated under three missing-data scenarios: random point missing, continuous interval missing, and entire-curve omission. For the entire-curve case, a compromise model structure is adopted. Out-of-fold predictions are used for per-class error diagnosis, and TreeSHAP is applied to interpret the model’s decision rationale. The results show that XGBoost, using native missing-value handling and no class weighting, achieves the most balanced cross-well performance, with an accuracy of 0.687, a Macro-F1 of 0.328, and an MCC of 0.430. Introducing 30% random missing points reduced Macro-F1 by 36.98% and MCC by 39.63%, indicating that class-balance metrics are more sensitive than overall accuracy. Considerable confusion exists between the sandstone–shale transitional lithology and pure shale; 67.02% of true transitional samples are misclassified as shale. SHAP analysis of class-wise contributions reveals that RDEP, RHOB, and DTC are the most important logging curves overall, and feature dependence varies significantly among different lithologies. These findings can serve as a reference for batch preliminary lithology screening, logging quality control, and manual review when logging data are incomplete.
Downloads
References
[1] Busch, J. M., Fortney, W. G., & Berry, L. N. (1987). Determination of lithology from well logs by statistical analysis. SPE Formation Evaluation, 2(4), 412–418.
[2] Zhu, X., Zhang, H., Ren, Q., et al. (2024). A review on intelligent recognition with logging data: Tasks, current status and challenges. Surveys in Geophysics, 45(5), 1493–1526. https://doi.org/ 10.1007 /s 10712-024-09862-9.
[3] Zou, W. B. (2020). Artificial intelligence research status and applications in well logging. Well Logging Technology, 44(4), 323–328.
[4] Wang, Z. Z., Zhang, C. L., & Gao, S. C. (2017). Identification of complex carbonate lithology using decision tree method: A case study of Sudong 41 33 block in Sulige Gas Field. Petroleum Geology and Recovery Efficiency, 24(6), 25–33.
[5] Liu, H., Wu, Y., Cao, Y., et al. (2020). Well logging based lithology identification model establishment under data drift: A transfer learning method. Sensors, 20(13), 3643. https:// doi.org/ 10. 3390/ s 20133643.
[6] Hall, B. (2016). Facies classification using machine learning. The Leading Edge, 35(10), 906–909. https://doi.org/10.1190/tle35100906.1.
[7] Halotel, J., Demyanov, V., & Gardiner, A. (2020). Value of geologically derived features in machine learning facies classification. Mathematical Geosciences, 52(1), 5–29. https://doi.org/ 10.1007/ s11004-019-09820-3.
[8] Dev, V. A., & Eden, M. R. (2019). Formation lithology classification using scalable gradient boosted decision trees. Computers & Chemical Engineering, 128, 392–404. https://doi.org/ 10. 1016/ j. compchemeng. 2019.06.014.
[9] Bressan, T. S., Kehl de Souza, M., Girelli, T. J., et al. (2020). Evaluation of machine learning methods for lithology classification using geophysical data. Computers & Geosciences, 139, 104475. https://doi.org/10.1016/j.cageo.2020.104475.
[10] Zhou, K., Zhang, J., Ren, Y., et al. (2020). A gradient boosting decision tree algorithm combining synthetic minority oversampling technique for lithology identification. Geophysics, 85(4), WA147–WA158. https://doi.org/10.1190/geo2019-0598.1.
[11] Merembayev, T., Kurmangaliyev, D., Bekbauov, B., et al. (2021). A comparison of machine learning algorithms in predicting lithofacies: Case studies from Norway and Kazakhstan. Energies, 14(7), 1896. https://doi.org/10.3390/en14071896.
[12] Zhu, L., Li, et al. (2018). Intelligent logging lithological interpretation with convolution neural networks. Petrophysics, 59(6), 799–810.
[13] Imamverdiyev, Y., & Sukhostat, L. (2019). Lithological facies classification using deep convolutional neural network. Journal of Petroleum Science and Engineering, 174, 216–228. https://doi.org/10. 1016/j. petrol.2018.11.060.
[14] Liu, X., Shao, G., Liu, Y., et al. (2022). Deep classified autoencoder for lithofacies identification. IEEE Transactions on Geoscience and Remote Sensing, 60, 1–14. https://doi.org/ 10.1109/ TGRS. 2022. 3141479.
[15] Feng, R. (2021). Uncertainty analysis in well log classification by Bayesian long short-term memory networks. Journal of Petroleum Science and Engineering, 205, 108816. https://doi. org/ 10. 1016/ j. petrol.2021.108816.
[16] Zeng, L., Ren, W., & Shan, L. (2020). Attention-based bidirectional gated recurrent unit neural networks for well logs prediction and lithology identification. Neurocomputing, 414, 153–171. https:// doi.org/10.1016/j.neucom.2020.07.060.
[17] Santos, D. T. dos, Roisenberg, M., & Nascimento, M. dos S. (2022). Deep recurrent neural networks approach to sedimentary facies classification using well logs. IEEE Geoscience and Remote Sensing Letters, 19. https://doi.org/10.1109/LGRS.2021.3131757.
[18] Bormann, P., Aursand, P., Dilib, F., et al. (2020). FORCE 2020 well log and lithofacies dataset for machine learning competition. Zenodo. https://zenodo.org/record/4351155.
[19] Ellis, D. V., & Singer, J. M. (2007). Well logging for earth scientists. Springer Netherlands. https:// doi. org/10.1007/978-1-4020-5514-6.
[20] Al-Mudhafar, W. J. (2017). Integrating well log interpretations for lithofacies classification and permeability modeling through advanced machine learning algorithms. Journal of Petroleum Exploration and Production Technology, 7(4), 1023–1033. https://doi.org/10.1007/s13202-017-0368-1.
[21] Feng, R. (2021). Uncertainty analysis in well log classification by Bayesian long short-term memory networks. Journal of Petroleum Science and Engineering, 205, 108816. https://doi.org/10.1016/j. petrol. 2021.108816.
[22] Zhang, J., He, Y., Zhang, Y., et al. (2022). Well-logging-based lithology classification using machine learning methods for high-quality reservoir identification: A case study of Baikouquan Formation in Mahu Area of Junggar Basin, NW China. Energies, 15(10), 3675. https://doi.org/10.3390/ en 15 103675.
[23] Sun, Y., Pang, S., Zhao, Z., et al. (2024). Interpretable SHAP model combining meta-learning and vision transformer for lithology classification using limited and unbalanced drilling data in well logging. Natural Resources Research, 33(6), 2545–2565. https://doi.org/10.1007/s11053-024-09724-1.
[24] Li, N., Xu, B. S., Wu, H. L., Feng, Z., Li, Y. S., Wang, K. W., Liu, P. (2021). Application status and prospects of artificial intelligence in well logging and formation evaluation. Acta Petrolei Sinica, 42(4), 508–522.
[25] Carvalho, H. M. O., Daniel, H., Korenchendler, A., et al. (2026). Lithology classification based on well log data: A benchmark for machine learning models. Mathematical Geosciences.
[26] Jiang, C., Zhang, D., & Chen, S. (2021). Lithology identification from well-log curves via neural networks with additional geologic constraint. Geophysics, 86(5), IM85–IM100. https://doi.org/ 10. 1190/ geo2020-0472.1.
[27] Xie, D., Liu, Z., Wang, F., et al. (2024). A transformer and LSTM-based approach for blind well lithology prediction. Symmetry, 16(5), 616. https://doi.org/10.3390/sym16050616.
[28] Wang, J., & Cao, J. (2023). A lithology identification approach using well logs data and convolutional long short-term memory networks. IEEE Geoscience and Remote Sensing Letters, 20, 1–5. https: // doi. org/10.1109/LGRS.2023.3241318.
[29] Cao, Z. M., Liu, P. C., Han, J., et al. (2025). A lithology identification method based on meta-learning using multi-view resampled heterogeneous representative features from well logs. Geophysical Prospecting for Petroleum, 64(3), 575–586.
[30] Shao, X., Zhang, P., Yan, S., et al. (2025). A multi-model fusion network for enhanced blind well lithology prediction. Processes, 13(1), 278. https://doi.org/10.3390/pr13010278.
[31] Carvalho, H. M. O., Daniel, H., Korenchendler, A., et al. (2026). Lithology classification based on well log data: A benchmark for machine learning models. Mathematical Geosciences.
[32] Dong, S., Wang, L., Xu, T., et al. (2026). Improved recurrent neural network for complex lithology identification in gas reservoirs. Petroleum Science and Technology, 44(16), 2673–2694. https:// doi. org/10.1080/10916466.2026.2462118.
[33] Gama, P. H. T., Faria, J., Sena, J., et al. (2025). Imputation in well log data: A benchmark for machine learning methods. Computers & Geosciences, 196, 105789. https://doi.org/10. 1016/j. cageo. 2025. 105789.
[34] Garini, S. A., Shiddiqi, A. M., & Utama, W. (2025). Filling-well: An effective technique to handle incomplete well-log data for lithology classification using machine learning algorithms. MethodsX, 14, 103127. https://doi.org/10.1016/j.mex.2025.103127.
[35] Pang, Q., Chen, C., Li, W., et al. (2026). Lithology identification from missing well log data: a multi-constraint guided self-supervised diffusion framework. Advanced Engineering Informatics, 74, 104697. https://doi.org/10.1016/j.aei.2026.104697.
[36] Zhu, X., Zhang, H., Ren, Q., et al. (2023). An automatic identification method of imbalanced lithology based on Deep Forest and K-means SMOTE. Geoenergy Science and Engineering, 224, 211595. https: // doi.org/10.1016/j.geoen.2023.211595.
[37] IEEE. (2026). Causally informed domain-guided machine learning for imbalanced lithofacies classification from well logs. IEEE Xplore. [2026-07-30].
[38] Ekmekcioğlu, Ö, Koc, K., Özger, M., et al. (2022). Exploring the additional value of class imbalance distributions on interpretable flash flood susceptibility prediction in the Black Warrior River basin, Alabama, United States. Journal of Hydrology, 610, 127877. https://doi.org/10. 1016/j. jhydrol. 2022. 127877.
[39] Lundberg, S., & Lee, S. I. (2017). A unified approach to interpreting model predictions. arXiv preprint arXiv:1705.07874. https://doi.org/10.48550/arXiv.1705.07874.
[40] Fetni, Y., Ameur-Zaimeche, O., Wood, D. A., et al. (2025). Explainable machine-learning workflow to improve real-time lithofacies identification using mudlogging data from a heterogeneous clastic reservoir sequence. Computational Geosciences, 29(6), 56. https://doi.org/10.1007/s10596-025-10367-1.
[41] Sun, Y., Pang, S., Li, H., et al. (2025). Enhanced lithology classification using an interpretable SHAP model integrating semi-supervised contrastive learning and transformer with well logging data. Natural Resources Research, 34(2), 785–813. https://doi.org/10.1007/s11053-025-09916-9.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Scientific Journal of Technology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
