【摘要】 在 k-mains聚类分析方法的基础上 ,提出一种新的聚类方法 ,将其与最小二乘法相结合 ,不仅提高了 RBFNN模型的精度和运算速度 ,而且可以动态的决定 RBFNN的结构。在此基础上 ,提出了对 RBF中心建模的方法 ,并将其应用于油田累计产量系统的建模和预测 ,结果表明这些方法是有效而适用的。更多还原

【Abstract】 On the basis of k mains clustering analysis method ,the paper puts forward a new clustering method and combines it with least squares algorithm.This method not only improves the precision and the operational speed of RBFNN model but also can decide the structure of RBFNN dynamically.On this basis,this paper also puts forward the method of RBF center modeling and applys it to the modeling and forecast of an oil field accumulative output system.And the outcome indicates that these methods are effective and applicable.更多还原