Mathematics and Statistics Colloquium

Calvin University

Joyce Chew
Calvin University

November 06, 2025
3:05pm   •   North Hall 276

Topic Modeling of Text Data with Nonnegative Matrix Factorization

When humans read a collection of text documents, one natural way to describe a document in the collection is to identify its main topics. The technique of topic modeling aims to mathematically and computationally mimic this process. In this talk, I will introduce nonnegative matrix factorization (NMF) as a method for learning topics from large collections of documents. I will then briefly survey perspectives on how to compute these factorizations. Finally, I will show how the framework of NMF can be augmented and adapted for specific applications.