基于RBF网络曲线拟合的研究

Research of curve fitting based on RBF network

曲线拟合问题是求得自变量与因变量之间的逼近函数关系表达式,在MATLAB中可通过编写RBF网络实现曲线拟合,但是样本数据和参数的选择是值得研究的问题,用不同的样本数据与不同的参数进行曲线拟合,并把拟合出的效果进行比较,从而得出怎样选择数据样本和参数。

The key to curve fitting is to obtain the function expression between independent variable and the dependent variable .Curve fitting can be obtained in written RBF networks in matlab ,but the choice of sample data and parameters is a problem worthy of study . The different sample data with different parameters is given ,and the curve fitting results will be compared to determine how to select data samples and parameters .