vsp - Vintage Sparse PCA for Semi-Parametric Factor Analysis
Provides fast spectral estimation of latent factors in random dot product graphs using the vsp estimator. Under mild assumptions, the vsp estimator is consistent for (degree-corrected) stochastic blockmodels, (degree-corrected) mixed-membership stochastic blockmodels, and degree-corrected overlapping stochastic blockmodels.
Last updated 13 days ago
6.35 score 25 stars 20 scripts 132 downloadsfastRG - Sample Generalized Random Dot Product Graphs in Linear Time
Samples generalized random product graphs, a generalization of a broad class of network models. Given matrices X, S, and Y with with non-negative entries, samples a matrix with expectation X S Y^T and independent Poisson or Bernoulli entries using the fastRG algorithm of Rohe et al. (2017) <https://www.jmlr.org/papers/v19/17-128.html>. The algorithm first samples the number of edges and then puts them down one-by-one. As a result it is O(m) where m is the number of edges, a dramatic improvement over element-wise algorithms that which require O(n^2) operations to sample a random graph, where n is the number of nodes.
Last updated 3 months ago
adjacency-matrixgraph-samplinglatent-factors
4.52 score 5 stars 22 scripts 219 downloadsfastadi - Self-Tuning Data Adaptive Matrix Imputation
Implements the AdaptiveImpute matrix completion algorithm of 'Intelligent Initialization and Adaptive Thresholding for Iterative Matrix Completion', <https://amstat.tandfonline.com/doi/abs/10.1080/10618600.2018.1518238>. AdaptiveImpute is useful for embedding sparsely observed matrices, often out performs competing matrix completion algorithms, and self-tunes its hyperparameter, making usage easy.
Last updated 5 months ago
4.43 score 9 stars 6 scripts 198 downloadsLRMF3 - Low Rank Matrix Factorization S3 Objects
Provides S3 classes to represent low rank matrix decompositions.
Last updated 3 years ago
matrix-factorizationsingular-value-decomposition
3.78 score 2 stars 2 packages 9 scripts 106 downloadsgdim - Estimate Graph Dimension using Cross-Validated Eigenvalues
Cross-validated eigenvalues are estimated by splitting a graph into two parts, the training and the test graph. The training graph is used to estimate eigenvectors, and the test graph is used to evaluate the correlation between the training eigenvectors and the eigenvectors of the test graph. The correlations follow a simple central limit theorem that can be used to estimate graph dimension via hypothesis testing, see Chen et al. (2021) <arXiv:2108.03336> for details.
Last updated 1 years ago
3.48 score 6 stars 6 scripts 143 downloadssparseLRMatrix - Represent and Use Sparse + Low Rank Matrices
Provides an S4 class for representing and interacting with sparse plus rank matrices. At the moment the implementation is quite spare, but the plan is eventually subclass Matrix objects.
Last updated 4 years ago
3.48 score 1 stars 2 packages 2 scripts 123 downloadsinvertiforms - Invertible Transforms for Matrices
Provides composable invertible transforms for (sparse) matrices.
Last updated 2 years ago
3.18 score 1 stars 1 packages 4 scripts 188 downloads