回归函数断点的小波识别方法

Wavelets Detection of Jump Points in Regression Model

断点识别是统计上研究的一类重要问题 ,利用小波方法作误差独立模型的回归函数的断点识别问题 ,给出了断点个数、位置及跃度的相合估计 ,结果表明小波方法比传统方法更有效 。

Jump points detection is a important problem in statistics.Proposed a method for the Detction of jump points in white noise model by wavelets.And the estimation of the number,locations and magnitudes of jumps are gotten.The results suggest that this method is more effective than traditional method.