Continuous Semi-Supervised Nonnegative Matrix Factorization

TL;DR

This paper introduces a continuous semi-supervised nonnegative matrix factorization method that improves topic detection and regression performance simultaneously, enhancing interpretability and predictive accuracy.

Contribution

It presents a novel integration of NMF with regression on a continuous response, combining topic modeling with supervised learning in a unified framework.

Findings

Outperforms traditional post-hoc regression methods

Enhances interpretability of topic models

Provides better predictive accuracy on continuous responses

Abstract

Nonnegative matrix factorization can be used to automatically detect topics within a corpus in an unsupervised fashion. The technique amounts to an approximation of a nonnegative matrix as the product of two nonnegative matrices of lower rank. In this paper, we show this factorization can be combined with regression on a continuous response variable. In practice, the method performs better than regression done after topics are identified and retrains interpretability.