Study of Noise Reduction about Economic Time Series Data Based on ICA
Cheng Liu, Jian-kang Liu, Guo-zhu He · 2010
This paper raised a new method about noise-reduction for 1D economic time series data based on ICA, making use of high rank statistical peculiarity of ICA, because of a large number of noises contained in economic series data. We expand 1D signal to many dims, drawing into invented noise components, and separate blind source. Compared to adapt-self filter, this method is excel enter. Authentic proof making effect clearly.