| 刘秀峰,朵德英,王斌.新疆大河沿地区金矿成矿规律及预测方法[J].矿产勘查,2025,16(9):2111-2119 |
| 新疆大河沿地区金矿成矿规律及预测方法 |
| Metallogenic laws and prediction methods of gold deposits in the Daheyan Area of Xinjiang |
| 投稿时间：2023-08-08 |
| DOI：10.20008/j.kckc.202509004 |
| 中文关键词: 金矿成矿规律 成矿预测 SVM算法 随机森林 大河沿地区 新疆 |
| 英文关键词: metallogenic laws of gold deposits metallogenic prediction SVM algorithm random forest Da-heyan Area Xinjiang |
| 基金项目:本文受青海省地质矿产勘查开发局及青海省第一地质勘查院项目(青地矿科【2021】80号)资助。 |
|
| 摘要点击次数: 386 |
| 全文下载次数: 205 |
| 中文摘要: |
| 常规的金矿成矿规律及成矿预测方法，对成矿空间的识别缺乏精确度，导致金矿成矿预测准确性较差。基于此，本文提出新疆大河沿地区金矿成矿规律及成矿预测方法。首先，对收集的数据进行清洗和归一化处理；其次，基于 SVM算法分析成矿控制因素，获得成矿重要特征，由此实现成矿规律分析，为识别可能的矿化区域奠定基础，提高金矿成矿预测成功率；然后采用小波分解方法提取潜在的成矿信息，并对空间权重系数进行计算，再结合已有的成矿标签数据建立识别函数，实现成矿有利部位识别；最后，基于识别结果，使用随机森林方法，实现新疆大河沿地区金矿成矿预测。结果表明，使用本文设计的方法，成矿预测的 AUC值达到 0.9以上，能够对成矿进行准确的预测，具有较好的应用价值。 |
| 英文摘要: |
| The conventional metallogenic regularity and metallogenic prediction methods of gold mines lack accu-racy in identifying the metallogenic space, resulting in poor accuracy of metallogenic prediction of gold mines. Based onthis, this paper proposes the metallogenic regularity and metallogenic prediction methods of gold deposits in the Daheyanarea of Xinjiang. First, clean and normalize the collected data; secondly, based on the SVM algorithm to analyze the con-trolling factors of mineralization and obtain the important characteristics of mineralization, the analysis of mineralizationlaws can be achieved, laying the foundation for identifying possible mineralization areas and improving the success rateof gold mineralization prediction. Then, the wavelet decomposition method is adopted to extract the potential mineraliza-tion information, and the spatial weight coefficient is calculated. Combined with the existing mineralization label data, an identification function is established to achieve the identification of favorable mineralization locations. Finally, based onthe identification results, the random forest method was used to achieve the prediction of gold mineralization in the Da-heyan area of Xinjiang. The results show that by using the method designed in this paper, the AUC value of mineraliza-tion prediction reaches above 0.9, which can accurately predict mineralization and has good application value. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |