基于多层递阶回归分析的轧钢煤气用量预测

以某钢铁企业为背景,基于煤气用户的历史数据,通过多层递阶回归分析建立相应的消耗预报模型,从而对煤气用量进行预测。首先把统计样本中的各个高相关因子作为回归变量进行线性回归处理,然后以回归系数与预报因子的乘积作为对修正量来进行多层递阶预报。这种多层递阶与回归分析方法,既能较好地体现高相关因子在预报模型中的重要作用,又具有较强的适应性,可提高预报精度。实际工业应用证明了方法的有效性。

Prediction of Gas Consumption in the Rolling Mill Process Based on Multi-level Recursive Regression Method

A kind of new approach, called the multi-level recursive regression method, is used in the prediction problem on gas consumption of rolling mill. The variables in the original mode are taken as regression variables which function as linear regression, and then the product of the regression coefficient and the regression variants which functions as linear regression, and then the product of the regression coefficient and the predictive factor is taken as the revised variant of the original predictive factor. This method reflects the important function of the related factor in the prediction mode as well as the strong adaptability to the system of sequential variance. Applied result demonstrate that this method is satisfactory.