libNMF

The library contains implementations of different optimization and regularization variants of non-negative matrix factorization.

List of algorithms implemented:

1. Multiplicative Update Rule (nmf.py)
2. Alternating Least Squares NMF (alsnmf.py)
3. Graph Regularized NMF (gnmf.py)
4. Probabilistc NMF (pnmf.py)
5. Kernel NMF (knmf.py)
6. Chambolle-Pock based first-order primal dual algo (fpdnmf.py)

Setup:

To get the project's source code, clone the github repository:

$ git clone https://github.com/satwik77/libnmf.git

Install VirtualEnv using the following (optional):

$ [sudo] pip install virtualenv

Create and activate your virtual environment (optional):

$ virtualenv venv
$ source venv/bin/activate

Install all the required packages:

$ pip install -r requirements.txt

Install the library by running the following command from the root directory of the repository:

$ python setup.py install

Usage:

>>> import numpy as np

>>> # For Graph Regularized NMF
>>> from libnmf.gnmf import GNMF
>>> X = np.random.random((10,10))
>>> gnmf= GNMF(X, rank=4)
>>> gnmf.compute_factors(max_iter= 20, lmd= 0.3, weight_type='heat-kernel', param= 0.4)

>>> # For first-order primal-dual algo
>>> from libnmf.fpdnmf import FPDNMF
>>> fpdnmf= FPDNMF(X, rank=4)
>>> fpdnmf.compute_factors(max_iter=30, nditer=5)
>>> #print fpdnmf.W, fpdnmf.H, fpdnmf.div_error

Refer to examples/Simple-Usage.ipynb for more on usage.

References

• [1] Lee, D. D., & Seung, H. S. (2001). Algorithms for non-negative matrix factorization. In Advances in neural information processing systems (pp. 556-562). Paper

• [2] Lee, D. D. and Seung, H. S. (1999), Learning the Parts of Objects by Non-negative Matrix Factorization, Nature 401(6755), 788-799. Paper

• [3] Cai, D., He, X., Han, J., & Huang, T. S. (2011). Graph regularized nonnegative matrix factorization for data representation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(8), 1548-1560. Paper

• [4] Bayar, B., Bouaynaya, N., & Shterenberg, R. (2014). Probabilistic non-negative matrix factorization: theory and application to microarray data analysis. Journal of bioinformatics and computational biology, 12(01), 1450001. Paper

• [5] Zhang, D., Zhou, Z. H., & Chen, S. (2006). Non-negative matrix factorization on kernels. PRICAI 2006: Trends in Artificial Intelligence, 404-412. Paper

• [6] Yanez, Felipe, and Francis Bach. "Primal-dual algorithms for non-negative matrix factorization with the Kullback-Leibler divergence." In Acoustics, Speech and Signal Processing (ICASSP), 2017 IEEE International Conference on, pp. 2257-2261. IEEE, 2017. Paper