中文摘要：

在应用经验模态分解（EMD）处理数据的时候。端点效应成为影响该方法精度的主要因素，即在“筛”的过程中上下包络在数据序列的两端会出现发散现象。端点效应会增加一些虚假成分，信号的总能量也随之增加，本文提出可以计算EMD后信号的总能量，来评估端点效应的影响程度。为了抑缶喘点效应，本文提出一种新的解决方法，即在信号序列上加窗函数，仿真结果表明，窗函数法有明显的效果。

英文摘要：

The end effect is a key shortcoming of the Empirical Mode Decomposition. On the sifting process, the upper and lower envelop can have large swing. The end effect provides some fake component, and the energy of the signal is increased. An index based on energy has been developed for evaluating the end effect. The end effect can be reduced by multiplying a window function with the signal. The experiment result show that the method is feasible.