论文部分内容阅读
将支持向量机(SVM)和遗传算法(GA)集成应用到矿体品位插值问题中,利用遗传算法全局搜索的优势对支持向量机的三个关键参数——惩罚系数C、不敏感系数ε和核函数参数σ进行寻优,克服单纯支持向量机法中依靠经验确定参数的局限性.将优化参数代入到支持向量机中进行迭代训练,得到基于遗传算法参数优化的支持向量机(GA-SVM)矿体品位插值模型.以国内典型矿山的实际勘探数据为例,通过该品位插值模型计算结果与传统插值方法计算结果和矿山生产实际数据的对比分析,验证了其可行性和有效性.
The integration of support vector machine (SVM) and genetic algorithm (GA) is applied to the problem of orebody grade interpolation. The advantages of global search of genetic algorithm are used to evaluate the three key parameters of SVM - penalty coefficient C, insensitive coefficient ε and Kernel function parameter σ to overcome the limitations of the simple support vector machine (SVM) method, which is empirically determined. The optimization parameters are substituted into the support vector machine for iterative training, and the GA-SVM ) Orebody grade interpolation model, taking the actual exploration data of typical domestic mines as an example, the feasibility and effectiveness of the model are verified through the comparison and analysis of the calculated result of the grade interpolation model with the calculation results of traditional interpolation methods and the actual data of mine production.