“Looking through the Noise:” Enhancing Interpretability of Factor Analysis through Sparsity by Trishia Cueto ’24 and Kent Barbir ’24, Wednesday November 29, 1:10pm – 1:50pm, North Science Building 015, Wachenheim
Abstract: Dimension reduction and estimating hard-to-measure variables are common issues in statistical analyses. Factor analysis techniques attempt to address both. In this talk, we will start by giving a brief introduction to the traditional exploratory factor analysis. We will then discuss the challenges and extension of it by involving sparsity in factor loadings as a way to enhance interpretability as well as incorporating the underlying group structure. We will also cover an application of this sparse factor analysis to estimating partisanship in the US Congress through voting patterns.