改进多元线性回归模型在某油田产量预测中的应用

Application of Improved Multivariate Linear Regression to Output Prediction of an Oilfield

为了预测油田的动态产量,根据实际工作经验选择一些与油田产量有关的因素作为建模变量,通过对多个变量的综合回归分析,确定影响油田产量的重要因素,并建立适合某油田年产量预测的多元线性回归模型。分析回归参数求解过程中的重要信息,并建立了年产油量的改进回归预测模型。同时,将两种预测模型应用于某油田产量预测,通过比较,可以看出改进回归模型形式更加简单,预测结果更准确。

In order to predict the dynamic output of an oilfield, some relevant factors with oil output are chosen on the basis of actual manufacturing experience. The forecast model of an oilfield is built according to the important factors influencing oilfield output which are obtained from synthetical regression analysis. The improved regression model for predicting annual output of an oilfield is built up by analyzing the statistic and the important information from the regression parameters. Meanwhile, the two forecast models are used to predict the output of an oilfield. The improved regression model is better by comparing the results.