基于支持向量机的曲线重建方法

Curve Reconstruction Based on Support Vector Machine

基于统计学习理论 ( SLT)的支持向量机 ( SVM)在高维空间中表示复杂函数是一种有效的通用方法 ,也是一种新的、很有发展前景的机器学习算法。文中简要介绍了基于支持向量机的理论 ,并在此基础上提出了一种基于支持向量机 ( SVM)的曲线重建算法 ,最后给出了实验 ,证明了该方法的有效性。

The existing method of neural network for function approaching in curve reconstruction suffers from the shortcoming of local minimum and slow convergence rate. A curve reconstruction algorithm based on SVM theory is presented. This algorithm can improve the standardization ability of its learning machine according to the principle of minimal structure risk. This algorithm avoids the iterative operation of BP learning algorithm, so its convergence rate is faster than that of the BP algorithm. The advantages of this algorithm are that it requires less samples to reconstruct a curve and its approaching precision is higher. It ensures that the extremum is global. Simulation result shows that this algorithm is effective.