No articles match
Clustering, Consensus, and Classification5 days ago
Why NMF for Clustering? | API Reference | Consensus NMF | Divisive Clustering | Factor Matching | Classification from Embeddings | Theory | Example 1: Cancer Subtype Discovery (Golub Leukemia) | Going Deeper: k = 5 Reveals Substructure | Example 2: Hierarchical Clustering with dclust (AML Methylation) | Splitting Hierarchy | Example 3: Classification from NMF Embeddings (AML Subtypes) | UMAP of Factor Embeddings | What's Next
Cross-Validation for NMF Rank Selection5 days ago
Motivation | API Reference | Cross-Validation via nmf() | mask Semantics | Theory | Speckled Holdout | Choosing $k$ | Overfitting Dynamics | Multiple Replicates | Worked Examples | Example 1: Recovering True Rank from Synthetic Data | Example 2: Overfitting Dynamics (Epoch Plot) | Example 3: Recommendation Data with mask = "zeros" | Example 4: Ecological Survey Data | Example 5: Effect of test_fraction | Next Steps
Factorization Graphs5 days ago
Why Factorization Graphs? | Building Blocks | Example 1: Graph ≡ Plain NMF | Example 2: Multi-Modal Shared Factorization | Bird Communities + Geography | Comparison: Shared vs. Separate Factorizations | Example 3: Deep NMF | Coarse → Fine in AML | The Connecting Matrix | Example 4: Target Regularization in Graphs | Example 5: Cross-Validation over Graph Parameters | Composable Architectures | Summary | What's Next
Getting Started with RcppML5 days ago
What Is RcppML? | Installation | Quick NMF Demo | Quick SVD Demo | Quick Cross-Validation Demo | Built-in Datasets | Where to Go Next | Core Techniques | Advanced Methods | Infrastructure
GPU-Accelerated Computing5 days ago
Motivation | Quick Start | Benchmarks on Real Data | Data Preparation | Results | Key Observations | Why GPU Gains Are Larger for CV and IRLS | API Reference | Resource Override Priority | GPU-Supported Features | GPU-Direct StreamPress Reading | Building the GPU Library | Requirements | Building from a source checkout | Library Search Path | GPU Memory | HPC / SLURM Usage | Troubleshooting | Key Takeaways
Guided NMF: Label-Informed Factorization5 days ago
Motivation | Setup: Hawaii Birds Dataset | Step 1: Baseline NMF | Step 2: Post-Hoc Refinement | Step 3: In-NMF Target Regularization | Comparison | Visualizing the Embeddings | UMAP Projections | Factor Composition by Island | Guidance on Choosing a Method | Removing Label-Specific Information (Batch Removal) | Comparing Enrichment vs Removal | UMAP: Enrichment vs Removal | Factor Composition: Enrichment vs Removal | Integration with Factor Graphs
NMF Fundamentals5 days ago
Motivation | API Reference | The nmf() Function | Factor Extraction and Inspection | Reconstruction and Loss | Model Comparison | Convergence Control | Theory and Algorithms | Objective | Alternating NNLS | Initialization | Non-uniqueness | Projective NMF | Worked Examples | Example 1: Recovering Known Factors from Synthetic Data | Example 2: Choosing Rank with Cross-Validation | Per-Iteration Convergence: Underfitting vs. Overfitting | Example 3: Methylation Factor Interpretation (AML Data) | Example 4: Convergence and Initialization Comparison | Next Steps
Regularization and Constraints5 days ago
Why Regularize? | API Reference | Sparsity Penalties | Structural Penalties | Box Constraints | Parameter Conventions | Theory | Example 1: L1 Sparsity Spectrum on Swimmers | Example 2: Angular Decorrelation on Face Data | Angular Penalty Sweep | L1 for Parts-Based Decomposition | Example 3: Graph Laplacian for Spatial Smoothness | What's Next
Statistical Distributions for NMF5 days ago
Motivation | API Reference | Distribution Selection in nmf() | Dispersion Control | Zero-Inflation | Diagnostic Functions | Theory | Variance-Mean Relationship | IRLS | Dispersion | Recommended Workflow | Information Criteria | Worked Examples | Example 1: Distribution Auto-Selection on Count Data | Example 2: Score Test Diagnostics | Example 3: Zero-Inflation Detection and Modeling | Next Steps
StreamPress: High-Performance Sparse Matrix I/O5 days ago
Motivation | API Reference | Writing Sparse Matrices | Reading Sparse Matrices | Inspecting Files | Dense Matrix Support | Theory | Columnar Storage | Compression Pipeline | Precision Hierarchy | Worked Examples | Example 1: Sparse Matrix Round-Trip | Example 2: Dense Matrix Round-Trip | Example 3: Precision Mode Comparison | Example 4: Real-World Dataset — pbmc3k Single-Cell RNA-seq | Example 5: Out-of-Core Streaming NMF (Conceptual) | When StreamPress Wins | Why SPZ Beats Raw CSC Even at 1 Thread | Parallel Scaling | The I/O Bandwidth Picture | Summary: When to Use SPZ | Thread Guidance | Key Takeaways
SVD and PCA5 days ago
Motivation | API Reference | svd(A, k, ...) | pca(A, k, ...) | Sparse PCA and Nonneg PCA | Inspection Methods | Theory | SVD Decomposition | PCA as Centered SVD | Sparse and Nonneg PCA | SVD Algorithms in RcppML | Algorithm Summary | Deflation (Sequential Rank-1 ALS) | Krylov (KSPR — Krylov-Seeded Projected Refinement) | Lanczos (Golub-Kahan Bidiagonalization) | IRLBA (Implicitly Restarted Lanczos Bidiagonalization) | Randomized SVD (Stochastic Projection) | Automatic Method Selection | Runtime Comparison | Tuning Hyperparameters | Worked Examples | Example 1: Eigenfaces from Olivetti | Example 2: Sparse PCA for Interpretable Loadings | Example 3: Nonneg SVD vs NMF | Example 4: Method Choice for Constrained Problems | Next Steps