Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 559 - Foundations of Data Science: The TRIPODS Experience
Type: Invited
Date/Time: Thursday, August 6, 2020 : 3:00 PM to 4:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #309315
Title: Riemannian Embedding Models for Relational Data
Author(s): Abel Rodriguez*
Companies: University of California, Santa Cruz
Keywords: Factor Analysis; Riemannian manyfolds; Topological Data Analysis

We describe a novel class of factor models for categorical data. Rather than embedding multivariate discrete response onto a low dimensional Euclidean space, these models embed them into a more general (prespecified) low-dimensional manifold. This approach endows the model with greater expressive power without sacrificing interpretability. We will particularly focus on models for spherical embeddings, which are readily motivated by applications in political science.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program