论文信息 - Comparative Study of Kurtosis-Based FastICA and RobustICA Using Spatial Smoothing Preprocessing

Comparative Study of Kurtosis-Based FastICA and RobustICA Using Spatial Smoothing Preprocessing

Independent Component Analysis (ICA) is expected to decompose the incident signals and detect each signal effectively and conveniently. ICA is a statistical technique utilizing only the property that incident signals are independent of each other. This paper discusses an improved method to introduce spatial smoothing preprocessing (SSP) into FastICA and RobustICA algorithms.

Nobuyoshi Kikuma | Kunio Sakakibara | Yuya Sato

[1] B. Erickson. Adaptive antenna technology aids wireless transmission , 1997 .

[2] Pierre Comon,et al. Robust Independent Component Analysis by Iterative Maximization of the Kurtosis Contrast With Algebraic Optimal Step Size , 2010, IEEE Transactions on Neural Networks.