Signal Dimension Estimation Using Principal Component Analysis

• Activity Number: 198 - Highlights from STAT
• Type: Topic-Contributed
• Date/Time: Tuesday, August 10, 2021 : 1:30 PM to 3:20 PM
• Sponsor: International Statistical Institute
• Author(s): Klaus Nordhausen* and Joni Virta and Sara Taskinen
• Companies: University of Jyväskylä and University of Turku and University of Jyväskylä
• Keywords: Principla Component Analysis; Dimension Reduction; Order Determination; Dependent Data; Blind Source Separation; Independent Component Analysis

Abstract

In this work, we develop inferential tools for determining the correct number of signals under a general noisy latent variable model, which includes as a special case, for example, the noisy independent component model. The problem is approached using hypothesis testing, and we provide both a large-sample test and several resampling-based alternatives assuming independent observations as well as under the assumption of serial dependence. Simulations reveal that our suggested approaches keep the desired levels and have good power.