probability oriented framework for separation and recognition of electroencephalogram

2015 Fiscal Year Final Research Report

• Project/Area Number: 26820152
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Measurement engineering
• Research Institution: University of Tsukuba
• Principal Investigator: Miyabe Shigeki 筑波大学, システム情報系, 助教 (50598745)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Keywords: 多チャネル信号処理 / 最尤推定 / 相関係数 / ウィシャート分布 / 仲上m分布

Outline of Final Research Achievements

We originally planned to study statistical multichannel signal processing for electroencephalogram, but we changed the research plan and focused on the fundamental analysis method, and studied estimation problem with complex multivariate probabilistic distribution. We proposed a statistical approach to estimate correlation coefficient between two complex signal channels under the condition where the phases are unobservable. As a byproduct of this research, we found a maximum likelihood parameter estimation algorithm for bivariate Nakagami-m distribution and Kibble's bivariate gamma distribution. We also proposed maximum likelihood estimation algorithm for Wishart distribution, which is a conjugate prior of multivariate distributions.

Free Research Field

統計的信号処理