Ensemble learning lithology prediction based on fast non-dominated sorting algorithm

Lithology identification plays a key role in reservoir characterization tasks of oil and gas exploration and engineering. Traditional lithology identification relies on professional knowledge and manual experience, which is laborintensive, time-consuming and affected by subjectivityand inconsistency of expert experience. A variety of supervised learning methods such as naive Bayes and support vector machine have been applied, but it is difficult to directly compare the differences between data sets and lithology classification schemes. In this study, an integrated learning model based on support vector machine, random forest, multi-layer perceptron and XGBoost is constructed. The fast non-dominated sorting optimization algorithm is used to optimize the hyper-parameters of the meta-classifier in the integrated learning model. Through experiments, the model is tested on specific lithology data samples, and the performance of the model is evaluated by indicators such as accuracy and recall rate. The experimental results show that the ensemble learning model optimized by hyperparameters has a significant improvement in the accuracy of lithology classification, which effectively overcomes the shortcomings of traditional methods, provides a more reliable and efficient solution for lithology identification, and helps to promote the development of lithology identification technology in the field of oil and gas exploration and engineering.