Constraint Independent Component Analysis Algorithms and Applications
Xianchuan Yu, Dan Hu, Jindong Xu · 2014
In this chapter, the CICA algorithm (constrained independent component analysis) was proposed to solve the connatural problems of ICA (independent component analysis), such as the instability of the amplitude and the nondeterminacy of the order. To eliminate the uncertainty of the ICA algorithm, we use kurtosis as a statistical measure to sort the independent components and incorporate a projection method to normalize the independent component amplitude to 1. , we applied CICA to voice signals and mineral prediction, and the results of this algorithm more closely match the actual situation.