一种基于负熵最大化的改进的独立分量分析快速算法

An Improved Fast ICA Algorithm Based on Negentropy Maximization

在深入分析FastICA算法的基础上,提出了一种基于牛顿迭代算法和最速下降法相结合的独立分量分析改进算法。仿真结果表明,改进的算法能有效提高收敛速度,避免原算法中收敛速度不均匀问题。

A new algorithm combined Newton iteration and gradient method is proposed on the basis of studying Fast ICA extensively. The simulation shows that using the improved algorithm convergence speed can be increased, and the unequal convergence speed in the original algorithm can be avoided.