Dear He Zhang, Non-negative Matrix Factorization has active life in research - hence it has future. Currently, many researches dealing with the problem of detecting and localizing objects in still, gray-scale images are using Non-negative Matrix Factorization.
Apart from image analysis, NMF is also used in Text Mining to extract concept-based factors from a text corpus (i.e., a collection of documents). There are a few published studies that demonstrate NMF's applicability to concept extraction. In my own research, however, I have found both Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) to be somewhat better in the way their results are understood by human users. Regarding NMF's future, it will depend on applications and software implementations. The image analysis and text mining applications of NMF a re important. The R package for NMF is a good computational resource. So yes, I see a future for NMF.
NMF can be used to separate Drum and Harmonic components from music, check the work from Björn Schuller at TU Munich
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining