论文部分内容阅读
鉴于参考独立分量分析定义了所谓的接近性度量函数和与之相关的不等式,并把它作为约束项引入到负熵对比度函数中,取得了很好的分离效果,但存在若阈值选取不当则算法不收敛的问题.提出一个改进算法,算法的优化函数为负熵对比度函数和参考独立分量分析算法中的接近性度量函数之积,巧妙地避开了这个难以确定的阈值参数.针对合成数据和实际ECG数据的仿真实验表明,改进算法收敛快、提取效果好.
Since the reference independent component analysis defines the so-called proximity measurement function and its associated inequalities and introduces it as a constraint item into the negative entropy contrast function, a good separation effect is achieved. However, if the threshold is not properly selected, the algorithm An improved algorithm is proposed.The optimization function of the algorithm is the product of the closeness measure function of the negative entropy contrast function and the reference independent component analysis algorithm and cleverly avoids this difficult to determine the threshold parameter.According to the synthetic data and Simulation results of the actual ECG data show that the improved algorithm converges fast and the extraction effect is good.