Description: Independent Component Analysis (ICA) is a kind of blind signal separation method developed gradually in the past two decades. It is a statistical method that aims to separate the source signals from each other from the mixed signals collected by the sensors, making the isolated source signals as independent as possible. It in speech recognition, signal processing such as telecommunications and medical signal processing has been widely used, has become fanaticism, processing, artificial neural network, such as a hot research topic in the field of study.
This paper briefly expounds the development, applications and status quo of ICA, the principle of ICA is discussed in detail and the implementation process, systematically introduces the several main ICA algorithm and the inner link between them, on the basis of this analyzed a FastICA a fast ICA algorithm.
The nonlinear spectral signal of matter can be seen as a mixed signal composed of multiple independent source signals, and these independent source signals can be regarded as the characteristic signals of the spectrum. In order to better understand the characteristics of spectral signal, this paper, by using independent component analysis of thought and method, proposed the FastICA algorithm is utilized to extract spectral signal characteristics, and the simulation experiment was carried out.
File list (Check if you may need any files):
Download