论文部分内容阅读
提出了一种对铜锍品位进行预测的新方法 ,即以采集的现场数据为基础 ,采用系统辨识动态地建立了AR(p)模型与三次指数平滑模型 .AR(p)模型要求数据对象是平稳时间序列 ,而三次指数平滑模型的数据对象具有随机性 ,考虑到铜锍品位的波动性 ,将 2种模型按最小二乘原理 ,以组合预测误差平方和为目标函数 ,通过使误差平方和极小化来确定 2种预测方法的最优加权系数 ,建立了一种新的组合模型 ,其预测误差最小 .结果表明 ,在当时数据条件下 ,AR(p)与指数平滑组合模型比AR(p)与指数平滑模型单独使用时精确度都要高 ,这对指导生产具有实用意义 .
A new method of predicting copper matte grade is proposed, that is, the AR (p) model and the cubic exponential smoothing model are dynamically established by using system identification based on the collected field data.The AR (p) model requires that the data object be The data objects of the three-exponential smoothing model are stochastic. Considering the volatility of copper matte grades, the two models are modeled as the objective function according to the principle of least squares and the square sum of forecasting errors. (P) and exponential smoothing model AR (p) and exponential smoothing model AR (p (subscript t)) under the condition of the data at the time, the optimal weighted coefficient of the two prediction methods is determined, p) and exponential smoothing model used alone when the accuracy is high, which is of practical significance to guide the production.