Selecting a Non-Negative Factorization Model for Statistical Inference on Time Series of Graphs

• Activity Number: 402
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Nam Lee*+ and Youngser Park and Carey Priebe and Michael Rosen and I-Cheng Wang
• Companies: and Johns Hopkins University and Johns Hopkins University and Johns Hopkins University and Johns Hopkins University
• Keywords: Model Selection ; Pattern Analysis ; Network Science

Abstract

While non-negative factorization is a popular tool for analyzing non-negative data, as a model selection technique, it can perform poorly when dealing with data with stochasticity. We develop model selection techniques that can be used to augment existing non-negative factorization algorithms, illustrating the performance of our algorithms via the application to problems of inference on time series of graphs. We motivate our approach with singular value decomposition, and illustrate our framework through numerical experiments using real and simulated data.