Topic: MaxICA with augmented genetic algorithm and application to EEG data
Speaker: Professor Zhang Chunming
Event date: 6/25/2019
Event time: 15:00 pm
Venue: Lecture Hall 1506, Building 9
Sponsor: School of Mathematics and Statistics, Institute of Science and Technology
Abstract: In many scientific disciplines, finding hidden influential factors behind observational data is essential but challenging. The majority of existing approachesrely on linear transformation, i.e., hidden components are linear combinations of original sources. Motivated from analyzing non-linear temporal signals in finance, genetics, and neuroscience, this paper proposes the “maximum independent component analysis (MaxICA),based on max-linear combinations of originalsources. In contrast to existing methods, MaxICA benefits from focusing on significant major source signals while filtering out ignorable signals. A major tool forparameter learning of MaxICA is the proposed ERD_GA algorithm, consisting ofthree schemes for the elite weighted sum selection, randomly combined cross over, and dynamic mutation. Extensive empirical evaluations demonstrate theeffectiveness of MaxICA in either extracting max-linearly combined essentialsources in many applications or supplying a better approximation for nonlinearlycombined source signals, such as EEG recordings analyzed in this paper.