Intuitionistic fuzzy set similarity measure for precise lithology identification of volcanic Rocks in the Liaohe Basin

Abstract The petrophysical characterization of igneous lithofacies in the eastern Liaohe Basin faces persistent challenges due to highly heterogeneous lithostratigraphic architectures and the scarcity of high-resolution borehole image logs essential for precise geological interpretation. Amid the big-data era, machine-learning algorithms have emerged as a promising avenue, offering novel perspectives. To enhance the accuracy of lithology recognition, this study proposes a lithology recognition algorithm grounded in the intuitionistic fuzzy set similarity measure. By leveraging 800 sets of traditional logging curve data, including natural gamma, density logging, corrected neutron porosity, acoustic time-difference, and deep lateral resistivity, the fundamental logging-response characteristics of various volcanic rocks are systematically compiled. Using 62.5% of the data as training samples, a lithology prediction model based on the intuitionistic fuzzy set similarity measure is constructed. Subsequently, 37.5% of the data are employed for prediction, and the model’s accuracy is verified. The results demonstrate that the average coincidence rate of this model for lithology prediction reaches 83.9%, enabling effective discrimination among the main volcanic rock types in the study area. This methodology not only provides a novel mathematical framework for lithology discrimination in volcanic terrains but also advances logging interpretation technology in complex geological settings through its interpretable IFS-based quantification.