Genetic Algorithm Optimized Light Gradient Boosting Machine for 3D Mineral Prospectivity Modeling of Cu Polymetallic Skarn-Type Mineralization, Xuancheng Area, Anhui Province, Eastern China

While geological data gathering expands and advances along with mineral exploration, there is still a need for more innovation and enrichment because there are few complete analytical tools for these data. The light gradient boosting machine (LightGBM) technique is used in this study to estimate the...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Natural resources research (New York, N.Y.) Jg. 32; H. 5; S. 1897 - 1916
Hauptverfasser: Li, He, Li, Xiaohui, Yuan, Feng, Zhang, Mingming, Li, Xiangling, Ge, Can, Wang, Zhiqiang, Guo, Dong, Lan, Xueyi, Tang, Minhui, Lu, Sanming
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.10.2023
Springer Nature B.V
Schlagworte:
Accuracy
Algorithms
Chemistry and Earth Sciences
China
Computer Science
Copper
Earth and Environmental Science
Earth Sciences
Fossil Fuels (incl. Carbon Capture)
Genetic algorithms
Geography
Global optimization
Machine learning
Mathematical Modeling and Industrial Mathematics
Mathematical models
Mineral exploration
Mineral Resources
Mineralization
Model accuracy
Modelling
Original Paper
Parameters
Physics
Statistics for Engineering
Support vector machines
Sustainable Development
Three dimensional models
China
Anhui China
GA–LightGBM
3D mineral prospectivity modeling
Ensemble learning
Skarn-type mineralization
ISSN:1520-7439, 1573-8981
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:While geological data gathering expands and advances along with mineral exploration, there is still a need for more innovation and enrichment because there are few complete analytical tools for these data. The light gradient boosting machine (LightGBM) technique is used in this study to estimate the prospectivity for minerals in three dimensions. However, because of the LightGBM model's large number of hyper-parameters, it is difficult to manually configure and alter model hyper-parameters, which has a substantial impact on the trained model's accuracy and dependability. Therefore, to optimize the LightGBM hyper-parameters, we use genetic algorithm (GA), which has the resilience and global optimization searchability in addressing difficult optimization problems. The GA–LightGBM algorithm for 3D mineral prospectivity modeling is the name we give to this combined approach. This study compares the GA–LightGBM algorithm against the GA-optimized support vector machine (SVM) and random forest (RF) algorithms in order to assess its applicability and superiority. The training set accuracy, test set accuracy, and Kappa coefficient values for the GA–LightGBM algorithm model were 0.9763, 0.9651, and 0.9453, respectively. The GA–LightGBM model's receiver operating characteristic curve was quite close to the upper left corner of the graph. The findings show that in terms of application and predictability, the GA–LightGBM ensemble learning approach performs better than the GA-optimized SVM and GA-optimized RF models. The outcomes of this study offer fresh perspectives for 3D mineral prospectivity modeling.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1520-7439
1573-8981
DOI:10.1007/s11053-023-10227-y