Stationary Subspace Analysis to alleviate non-stationarity in EEG Brain Computer Interface
Stationary Subspace Analysis (SSA, von Buenau et al. 2009) is an unsupervised learning method that finds subspaces in which data distributions stay invariant over time. It has been shown to be very useful for studying non-stationarities in various applications. I will introduce this technique and an example of SSA application to EEG Brain Computer Interface. Then, I will present the first SSA algorithm based on a full generative model of the data. This new derivation relates SSA to previous work on finding interesting subspaces from high-dimensional data in a similar way as the three easy routes to independent component analysis. I will also talk about research on semiparametric statistical approach to independent component analysis including my previous works. Finally, if I will have time, I will explain the on-going research topic towards combination of causal inference (Haufe et al. 2009) and source localization in collaboration with the members of Ishii lab.
- GCOE 知識グリッドコアセミナーとして開催．