The 19th In Silico Megabank Research Seminar will be held on Friday, December 14.

・Date/Time: December 14(Fri.) 17:00‐
・Venue: Seminar Room 303 (3F), Research Building No.3, Electrical Engineering and plied Physics
http://www.megabank.tohoku.ac.jp/english/info/access.html
・Title: Robust Sparse Regression Modeling
・Lecturer: Heewon Park (Chuo University)
・Abstract:
The L1 type of the regularization technique, which has been attracting attention recently, has the advantage in that presumption and the variable selection can be performed simultaneously by adding a L1 type of penalty to the likelihood function, and it is especially effective to analyze high-dimensional data including the genomic data. In this study, I introduce an effective algorithm for the analysis of the genomic data and propose a variety of robust sparse regression modeling techniques based on the L1 type of the regularization technique that is more effective than any existing research. As a result of verifying the proposal technique utilizing a numerical experiment, the effectiveness of it has become clear compared with existing techniques.

・Organizer: Masao Nagasaki