Oral Defense: "Different Estimation methods of the Basic Independent Component Analysis Model"

Speaker: Jennie An, Washington University in Saint Louis

Abstract: Inspired by classic cocktail-party problem, the basic Independent Component Analysis model is established. What differs Independent Component Analysis from other kinds of analysis is the non-Gaussian part of the data is considered here. Several approaches are given by finding the maximum non-Gaussianity which is measured by kurtosis, mutual information, and etc. 

     With each estimation, we need to optimize the functions of expectations of nonquadratic functions since it can help us to access the higher-order statistics of non-Gaussian part of the data. The choice of Fast Fixed-Point Independent Component Analysis algorithm is one of the most efficient batch algorithms to obtain the higher-order information which will be implemented through R here. 


Hosts: Jimin Ding & Jose Figueroa-Lopez