Sparse Logistic Singular Value Decomposition (SLSVD) for Binary Matrix Data 
We implement the Sparse Logistic Singular Value Decomposition (SLSVD) using the Majorization-Minimization (MM) and coordinate descent (CD) algorithms in this Python package.
Our package consists of three major components:
- Simulated binary data generation
- Sparse logistic SVD
- Metrics for evaluating estimations
There are two major functions in this package:
generate_data(n, d, rank, random_seed=123): This function generates random binary data points. It takes four parameters: n for the number of data points, d for the number of features, rank for the number of rank, and random_seed for ensuring reproducibility.
sparse_logistic_svd_coord(dat, lambdas=np.logspace(-2, 2, num=10), k=2, quiet=True, max_iters=100, conv_crit=1e-5, randstart=False, normalize=False, start_A=None, start_B=None, start_mu=None): This function performs Sparse Logistic Singular Value Decomposition (SLSVD) using Majorization-Minimization and Coordinate Descent algorithms.
n(integer): Number of data points.d(integer): Number of features.rank: Number of components.random_seed(integer): Random seed to ensure reproducibility.dat: Input data matrix.lambdas: Array of regularization parameters.k: Number of components.quiet: Boolean to suppress iteration printouts.max_iters: Maximum number of iterations.conv_crit: Convergence criterion.randstart: Boolean to use random initialization.normalize: Boolean to normalize the components.start_A: Initial value for matrix A.start_B: Initial value for matrix B.start_mu: Initial value for the mean vector.
SLSVD establishes itself as a valuable enhancement to the Python ecosystem. There is no function in the Python package scikit-learn has similar functionality, our implementation uses Majorization-Minimization and Coordinate Descent algorithms.
Make sure Miniconda or Anaconda is installed on your system
git clone [email protected]:andyzhangstat/SLSVD.git
cd SLSVD # Navigate to the cloned repository directory# Method 1: create Conda Environment from the environment.yml file
conda env create -f environment.yml # Create Conda environment
conda activate SLSVD # Activate the Conda environment
# Method 2: create Conda Environment
conda create --name SLSVD python=3.9 -y
conda activate SLSVDEnsure the Conda environment is activated (you should see (SLSVD) in the terminal prompt)
poetry install # Install the package using Poetry# Check line coverage
pytest --cov=SLSVD
# Check branch coverage
pytest --cov-branch --cov=SLSVD
poetry run pytest --cov-branch --cov=src
poetry run pytest --cov-branch --cov=SLSVD --cov-report html-
Environment Creation Issues: Ensure environment.yml is in the correct directory and you have the correct Conda version
-
Poetry Installation Issues: Verify Poetry is correctly installed in the Conda environment and your pyproject.toml file is properly configured
Use this package to find the optimized score and loading matrices of sparse logistic Singular Value Decomposition. In the following example, we generate a simulated data set with defined size first. By the Majorization-Minimization and Coordinate Descent algorithms, we obtain the optimized score and loading matrices. Finally, we visualize both the simulated data and fitted loadings in one figure.
Example usage:
>>> from slsvd.data_generation import generate_data
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> bin_mat, loadings, scores, diagonal=generate_data(n=200, d=100, rank=2, random_seed=123)
# Check shapes
>>> print("Binary Matrix Shape:", bin_mat.shape)
>>> print("Loadings Shape:", loadings.shape)
>>> print("Scores Shape:", scores.shape)
# Calculate dot product of scores
>>> scores_dot_product = np.dot(scores.T, scores)
>>> print("Dot Product of Scores:\n", scores_dot_product)
# Calculate dot product of loadings
>>> loadings_dot_product = np.dot(loadings.T, loadings)
>>> print("Dot Product of Loadings:\n", loadings_dot_product)Binary Matrix Shape: (200, 100)
Loadings Shape: (100, 2)
Scores Shape: (200, 2)
Dot Product of Scores:
array([[195.4146256 , 2.67535881],
[ 2.67535881, 200.14653178]])
Dot Product of Loadings:
array([[1., 0.],
[0., 1.]])
>>> plt.figure(figsize=(6, 9))
>>> colors = ['cyan', 'magenta']
>>> cmap = plt.matplotlib.colors.ListedColormap(colors, name='custom_cmap', N=2)
>>> plt.imshow(bin_mat, cmap=cmap, interpolation='nearest')
>>> cbar = plt.colorbar(ticks=[0.25, 0.75])
>>> cbar.ax.set_yticklabels(['0', '1'])
>>> plt.title('Heatmap of Simulated Binary Matrix')
>>> plt.xlabel('Feature')
>>> plt.ylabel('Sample')
>>> plt.tight_layout()
>>> plt.show()
>>> from slsvd.slsvd import sparse_logistic_svd_coord
>>> import numpy as np
>>> # Perform Sparse Logistic SVD
>>> mu, A, B, zeros, BICs = sparse_logistic_svd_coord(bin_mat, lambdas=np.logspace(-2, 1, num=10), k=2)
>>> # Calculate mean of mu
>>> print("Mean of mu:", np.mean(mu))
>>> # Calculate dot product of Scores
>>> print("Dot Product of Scores:\n", np.dot(A.T, A))
>>> # Calculate dot product of Loadings
>>> print("Dot Product of Loadings:\n", np.dot(B.T, B))Mean of mu: 0.052624279581212116
Dot Product of Scores:
array([[7672.61634966, 277.23466856],
[ 277.23466856, 3986.24113586]])
Dot Product of Loadings:
array([[1. , 0.00111067],
[0.00111067, 1. ]])
Online documentation is available readthedocs.
Publishing on TestPyPi and PyPi.
Interested in contributing? Check out the contributing guidelines. Please note that this project is released with a Code of Conduct. By contributing to this project, you agree to abide by its terms.
SLSVD was created by Andy Zhang. It is licensed under the terms of the MIT license.
SLSVD was created with cookiecutter and the py-pkgs-cookiecutter template.