基于加权支持向量机的选矿生产指标预报模型

Prediction Model for Production Index of Mineral Process Based on Weighted Support Vector Machine

在选矿生产中为了达到规定的目标,选矿工程师需要根据原矿的情况给各生产工序下达合理的生产任务,但由于选矿生产具有大滞后的特点,在选矿工程师决策下达执行前,如果能够对最终的生产结果进行预测则具有重要意义。以某选矿厂为实际背景,在分析了指标之间关系的基础上,采用加权支持向量机建立了相关指标的预报模型,并通过构造重要性函数的方法确定了支持向量机加权系数,最后利用该厂历史数据进行了仿真实验。

To achieve desired global production indices in mineral process, proper production tasks need to be distributed to every section by engineer based on the characteristics of ores. For the large delay in mineral process, it is reasonable to predict global production indices before the decision is set down. The prediction model using the weighted support vector machine was proposed considering the factors which had relations with the global production indices. The weighted coefficients of the support vector machine were determined by importance function. Simulation results show the efficiency of the prediction model using the industrial data.