Scholarlink

<p>Mineral prospectivity mapping (MPM) through the identification of prospective areas, by analyzing various exploration data and integrating them, plays a crucial role in reducing risk and improving decision-making in mineral exploration. However, this process is complex and faces many challenges due to the uncertainties inherent in the data and the various models used. In this study, our aim is to integrate several advanced methods to identify anomalies and use a new method to evaluate the performance of the developed models. To this end, five deep learning algorithms were employed for MPM, and their results were combined using a new method based on Bayesian statistics. This method was applied to five different models, resulting in a final composite model with a high level of confidence. The evaluation of results was performed using the Prediction Area plot (P-A plot). The final model demonstrated 6% higher accuracy compared to individual models and identified a smaller area as high-potential regions. Geologically, the results of the final model showed good alignment with microgranite, granodiorite to diorite, quartz diorite, and quartz monzodiorite units, indicating the success of this method in forward-looking modeling. The findings of this research suggest that combining models using this index can help reduce uncertainty and improve predictions in the identification of exploration targets, leading to more accurate decision-making and reduced exploration risks. This approach can be effectively applied in future exploration efforts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integration of deep learning models for mineral prospectivity mapping: a novel Bayesian index approach to reducing uncertainty in exploration

  • 期刊论文

  • 发表日期：

    2025年3月6日

查看全文
  • Zohre Hoseinzade,
  • Mojgan Shojaei,
  • Farkhondeh Khademi,
  • Ahmad Reza Mokhtari,
  • Mobin Saremi

摘要

Mineral prospectivity mapping (MPM) through the identification of prospective areas, by analyzing various exploration data and integrating them, plays a crucial role in reducing risk and improving decision-making in mineral exploration. However, this process is complex and faces many challenges due to the uncertainties inherent in the data and the various models used. In this study, our aim is to integrate several advanced methods to identify anomalies and use a new method to evaluate the performance of the developed models. To this end, five deep learning algorithms were employed for MPM, and their results were combined using a new method based on Bayesian statistics. This method was applied to five different models, resulting in a final composite model with a high level of confidence. The evaluation of results was performed using the Prediction Area plot (P-A plot). The final model demonstrated 6% higher accuracy compared to individual models and identified a smaller area as high-potential regions. Geologically, the results of the final model showed good alignment with microgranite, granodiorite to diorite, quartz diorite, and quartz monzodiorite units, indicating the success of this method in forward-looking modeling. The findings of this research suggest that combining models using this index can help reduce uncertainty and improve predictions in the identification of exploration targets, leading to more accurate decision-making and reduced exploration risks. This approach can be effectively applied in future exploration efforts.