Blind Source Separation Over Space

• Activity Number: 39 - Advances in Time Series: Statistics Meets Machine Learning
• Type: Invited
• Date/Time: Sunday, August 7, 2022 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): QIWEI YAO*
• Companies: London School of Economics
• Keywords: Eigen-analysis; Eigen-gap; High-dimensional random field; Mixing matrix; Spatial local covariance matrix

Abstract

We propose a new estimation method for the blind source separation model. The new estimation is based on an eigenanalysis of a positive definite matrix defined in terms of multiple spatial local covariance matrices, and, therefore, can handle moderately high-dimensional random fields. The consistency of the estimated mixing matrix is established with explicit error rates even when the eigen-gap decays to 0 (slowly). The proposed method is illustrated via both simulation and a real data example.