基于支持向量机的调剖措施效果预测模型研究

因调剖措施效果受多种因素的影响,很难在调剖措施效果和各个因素之间建立一种确定的关系。利用支持向量机方法,采用径向基核函数,建立了调剖措施效果与各个因素之间的一种关系模型,分别以无因次增油量、含水率的下降等评价指标建立不同的支持向量机模型,利用已实施的调剖措施作为学习样本来获得相应的模型,分别用于对无因次增油量、含水率的下降等指标进行预测。该模型的检验过程表明,检验样本的平均误差控制在10%以内,具有较高的准确率。

Model for Predicting the Efficiency of Profile Control Based on Support Vector Machine

Because the effect of profile control is influenced by various factors, thus it is hard to establish the relationship between the profile control and each factor. A relation model between the profile control effect and each factor is established by using radial kernel function based on support vector machine. Models based on different support vector machines are established by using the indexes, such as dimensionless oil increment and reduction of water content, the implemented profile control measures are used as a learning sample to build corresponding models, they are used to predict the indexes stated above. Inspection of the models indicates that the inspection error is controlled within 10 with higher precision.