Reproduce “Text Mining using Nonnegative Matrix Factorization and Latent Semantic Analysis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ali Hassani, Amir Iranmanesh, Najme Mansouri
A new method for text clustering, NMF-FR, combines Nonnegative Matrix Factorization (NMF)-based feature agglomeration with deterministic Spherical K-Means initialization. This approach enhances clustering performance across 12 text datasets and notably achieves near-zero standard deviation in results across multiple runs, ensuring high stability and reproducibility.
Key insights
- Fundamentally redefining the feature space through NMF-based term agglomeration, rather than just projection, can yield a representation more suitable for clustering.
- Achieving deterministic and stable clustering outcomes is critical for practical applications, and this can be realized through controlled initialization and feature engineering.
- Combining robust matrix factorization with a novel, deterministic initialization strategy effectively addresses both performance and reproducibility challenges in text clustering.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.