多金属矿产综合信息区域深部成矿预测--以临江市东港地区靶区预测为例

Deep Metallogenic Prediction of Polymetallic Mineral Comprehensive Information Area:Taking Target Area Prediction of Donggang District in Linjiang City as an Example

针对现有成矿预测方法存在预测精度低、效率不高的问题,提出多金属矿产综合信息区域深部成矿预测方法。采用地理信息系统(geographic information system,GIS)建立多金属矿综合信息库,数字化处理成矿信息。建立证据权模型,结合信息库,通过计算控矿地质因素的对比度,获取最优缓冲距离,完成区域深部成矿信息提取。基于此,利用遗传算法优化支持向量机参数,基于优化参数后的支持向量机完成成矿预测。实验结果表明:所提方法的最优缓冲距离计算准确率较高,支持向量机参数优化耗时短,且成矿预测精度在90%以上,不受数据状态的影响,预测度也优于传统方法,表明所提方法能够高效完成成矿预测。

In view of the low accuracy and efficiency of the existing metallogenic prediction methods,a deep metallogenic prediction method in the comprehensive information region of polymetallic minerals was proposed.geographic information system(GIS)was used to build the comprehensive information database of polymetallic ore and to digitize the metallogenic information.Based on the evidence weight model and the information database,the optimal buffer distance was obtained by calculating the contrast of ore control geological factors,and the information of regional deep mineralization was extracted.Genetic algorithm was used to optimize the parameters of SVM,and the prediction of mineralization was completed based on the optimized parameters of SVM.Results show that the accuracy of the optimal buffer distance calculation is high,the optimization time of SVM parameters is short,and the accuracy of metallogenic prediction reaches more than 90%,which is not affected by the data state,and the prediction degree is better than the traditional method,showing that the proposed method can complete metallogenic prediction efficiently.