[1]曹茂俊,巩维嘉,高志勇.基于 Stacking 集成学习的岩性识别研究[J].计算机技术与发展,2022,32(07):161-166.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 028]
　CAO Mao-jun,GONG Wei-jia,GAO Zhi-yong.Research on Lithology Identification Based on Stacking Integrated Learning[J].,2022,32(07):161-166.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 028]
点击复制

基于 Stacking 集成学习的岩性识别研究()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年07期
页码:
161-166
栏目:
应用前沿与综合
出版日期:
2022-07-10

文章信息/Info

Title:
Research on Lithology Identification Based on Stacking Integrated Learning
文章编号:
1673-629X(2022)07--0161-06
作者:
曹茂俊; 巩维嘉; 高志勇
东北石油大学 计算机与信息技术学院,黑龙江 大庆 163318
Author(s):
CAO Mao-jun; GONG Wei-jia; GAO Zhi-yong
School of Computer and Information Technology,Northeast Petroleum University,Daqing 163318,China
关键词:
岩性识别; K 近邻算法; 决策树; 支持向量机; 多层感知机; 改进集成学习
Keywords:
lithology identification;  K - nearest neighbor algorithm;  decision tree;  support vector machines;  multilayer perceptron;  improved integrated learning
分类号:
TP39
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 07. 028
摘要:
针对传统集成学习模型利用测井资料进行地层岩性识别效果不佳的问题,提出了一种改进的集成学习模型。 该模型使用 CART 决策树、K 近邻算法( KNN) 、支持向量机(SVM) 和多层感知机( MLP) 为基模型,逻辑回归( LR) 为元模型,使用 PCA 算法计算每个基模型的权重,并且将权重融入到第二层元模型的训练数据集中,从而给元模型提供了更多质量较高的训练数据,以此构建出一个精准的多层集成学习模型。 并且通过准确率、F1-Score 两个指标对该模型进行了评估。最后将改进后的集成分类模型应用在实际的井区数据中,实验表明改进后的模型相较于传统的 Stacking 集成模型准确率提升了 1. 85 个百分点,F1-Score 提升了 2 个百分点,与实际结果相比有较高的一致性,充分证明了改进后的集成分类模型的有效性。
Abstract:
Aiming at the problem that the traditional integrated learning model uses logging data to identify the formation lithology,? an improved integrated learning model is proposed. The model uses CART decision tree,K-nearest neighbor algorithm? ? ? ( KNN) ,support vector machine ( SVM) and multilayer perceptron ( MLP) as the base model,and logistic regression ( LR) as the meta-model. PCA algorithm is used to calculate the weight of each base model,which is integrated into the training data set of the second layer meta-model,there by providing more high-quality training data for the meta-model,so as to construct an accurate multi-layer ensemble learning model. And through two indicators of accuracy and F1-Score,the model is evaluated. Finally,the improved ensemble classification model is applied to the actual well area data. The experiment shows that the improved model prediction result is 1. 85 percentage points higher than the traditional Stacking ensemble model,and the F1 - Score is increased by 2 percentage points. Compared with the actual results,there is a higher consistency,which fully proves the effectiveness of the improved ensemble classification model.

相似文献/References:

[1]李昂儒,郑伟彦,赵京虎,等.基于分形理论的风电功率预测算法研究[J].计算机技术与发展,2021,31(03):191.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 033]
　LI Ang-ru,ZHENG Wei-yan,ZHAO Jing-hu,et al.Research on Wind Power Prediction Algorithm Based on Fractal Theory[J].,2021,31(07):191.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 033]
[2]张可佳,方佳佳,高楷程,等.基于神经网络的岩性物性参数计算方法研究[J].计算机技术与发展,2022,32(06):186.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 031]
　ZHANG Ke-jia,FANG Jia-jia*,GAO Kai-cheng,et al.Study on Calculation Method of Lithologic Physical Property Parameters Based on Neural Network[J].,2022,32(07):186.[doi:10. 3969 / j. issn. 1673-629X. 2022. 06. 031]
[3]刘重男,杨 洋,卢清心,等.低成本二维激光传感器室内移动建图方法研究[J].计算机技术与发展,2023,33(08):30.[doi:10. 3969 / j. issn. 1673-629X. 2023. 08. 005]
　LIU Zhong-nan,YANG Yang,LU Qing-xin,et al.Research on Indoor Mapping Based on Low-cost Two-dimensional Laser Sensor[J].,2023,33(07):30.[doi:10. 3969 / j. issn. 1673-629X. 2023. 08. 005]

更新日期/Last Update: 2022-07-10