支持向量机及其在提高采收率潜力预测中的应用

提高采收率潜力分析的基础是进行提高采收率方法的潜力预测 .建立提高采收率潜力预测模型从统计学习的角度来看 ,实质是属于函数逼近问题 .本文首次将统计学习理论及支持向量机方法引入提高采收率方法的潜力预测中 .根据 Vapnik结构风险最小化原则 ,应尽量提高学习机的泛化能力 ,即由有效的训练集样本得到的小的误差能够保证对独立的测试集仍保持小的误差 .在本文所用较少样本条件下 ,支持向量机方法能够兼顾模型的通用性和推广性 ,具有较好的应用前景 .研究中采用的是综合正交设计法、油藏数值模拟和经济评价等方法生成的理论样本集

Support Vector Machine and Its Application in Enhanced Oil Recovery Potentiality Prediction

The basis of Enhanced oil recovery (EOR) potentiality analysis is EOR potentiality prediction. In the view of statistical learning, to establish EOR predictive model substantially is to solve a function regression problem. In this paper, statistical learning theory and support vector machine method are introduced in EOR potentiality prediction for the first time. According to Vapnik structural risk minimization rule, it′s important to improve the generalization ability of learning machine. In other word, if there has small error for limited training sample set, then the error would keep small for independent testing sample set. In this paper, samples are very limited. SVM method can give attention to universal property and extension property at the same time, and it possesses good application prospect. A theory sample set generated by comprehensive application of quadrature design, reservoir simulation, and economical evaluation method is used in this research.