基于RF-PSO-SVM的测井岩性识别方法研究

朱斌; 赵军龙

文章摘要

基于RF-PSO-SVM的测井岩性识别方法研究

Research on Lithology Identification Method Based on RF-PSO-SVM

投稿时间：2025-05-13 修订日期：2025-06-06

DOI：

中文关键词: 岩性识别 RF算法 SVM算法 PSO算法 RF-PSO-SVM模型

英文关键词: lithology identification random forest algorithm support vector machine algorithm particle swarm optimization algorithm RF-PSO-SVM model

基金项目:陕西省自然科学基础研究计划(2019JM-359)

作者	单位	邮编
朱斌	西安石油大学	710065
赵军龙^*	西安石油大学	710065

摘要点击次数: 45

全文下载次数: 0

中文摘要:

针对常规岩性识别方法应用效果不理想的问题,提出一种基于RF-PSO-SVM的岩性识别模型。首先,通过RF算法中的OOB原则挑选出重要性高的测井参数；其次,基于PSO算法不同的粒子群数量寻优得到SVM模型最优参数组合；最后建立RF-PSO-SVM岩性识别模型对908条实验数据进行岩性预测,与PSO-SVM模型、SVM模型和RF模型相比较；识别准确率更高,达到95.97%,有效解决了常规岩性识别方法应用效果不理想的问题,为机器学习算法在岩性识别中的应用提供了一种优化思路。

英文摘要:

_: _{A lithology identification model based on RF-PSO-SVM is proposed to address the problem of unsatisfactory application effects of conventional lithology identification methods. Firstly, based on the Out-of-Bag (OOB) principle in the Random Forest (RF) algorithm, the logging parameters with high importance are selected. Secondly, by leveraging the powerful parameter optimization ability of the Particle Swarm Optimization (PSO) algorithm, the optimal parameter combination of the Support Vector Machine (SVM) model is found. Finally, the RF-PSO-SVM lithology identification model is established to conduct lithology prediction on 908 experimental data. When compared with the PSO-SVM model, the SVM model, and the RF model, it has a higher identification accuracy, reaching 95.97%. It effectively solves the problem of the unsatisfactory application effects of conventional lithology identification methods and provides an optimization idea for the application of machine learning algorithms in lithology identification.}

View Fulltext 查看/发表评论下载PDF阅读器

关闭