Package index • RcppML

Non-negative Matrix Factorization

Core NMF functions and model class

nmf(): Non-negative matrix factorization
nmf-class: nmf S4 Class
subset(<nmf>) `[`(<nmf>,<ANY>,<ANY>,<ANY>) head(<nmf>) show(<nmf>) dimnames(<nmf>) dim(<nmf>) t(<nmf>) sort(<nmf>) prod(<nmf>) `$`(<nmf>) coerce(<nmf>,<list>) `[[`(<nmf>) predict(<nmf>): nmf class methods
summary(<nmf>) plot(<nmfSummary>): Summarize NMF factors
biplot(<nmf>): Biplot for NMF factors
nnls(): Non-negative Least Squares Projection
compute_target(): Compute a Target Matrix for Guided NMF
refine(): Refine an NMF Model Using Label-Guided Correction

Singular Value Decomposition

SVD and PCA functions

svd(): Truncated SVD / PCA with constraints and regularization
`[`(<svd>,<ANY>,<ANY>,<ANY>) head(<svd>) show(<svd>) dim(<svd>) reconstruct() predict(<svd>) variance_explained(): svd S4 Class
pca(): PCA (centered SVD)

Model Evaluation & Diagnostics

Functions for evaluating and comparing NMF models

assess(): Assess Embedding Quality
as.data.frame(<nmf_assessment>): Convert assessment results to a one-row data frame
print(<nmf_assessment>): Print method for nmf_assessment objects
evaluate(): Evaluate an NMF model
sparsity(): Compute the sparsity of each NMF factor
align(): Align two NMF models
compare_nmf(): Compare Multiple NMF Models
cosine(): Cosine similarity

Distribution Selection

Automatic distribution and zero-inflation diagnostics

auto_nmf_distribution(): Auto-select NMF distribution
score_test_distribution(): Score-test distribution diagnostic
diagnose_zero_inflation(): Diagnose zero inflation
diagnose_dispersion(): Diagnose dispersion mode

Visualization

Plotting functions for NMF results

plot(<nmf>): Plot NMF Training History and Diagnostics
plot(<nmfCrossValidate>): Plot Cross-Validation Results

Consensus Clustering

Multi-replicate consensus NMF

consensus_nmf(): Consensus Clustering for NMF
plot(<consensus_nmf>): Plot Consensus Matrix Heatmap
summary(<consensus_nmf>): Summary for Consensus NMF

Clustering

NMF-based clustering algorithms

bipartition(): Bipartition a sample set
dclust(): Divisive clustering
plot(<dclust>): Plot divisive clustering hierarchy
bipartiteMatch(): Bipartite graph matching

Composable Factorization Graphs

Build complex factorization pipelines with factor_net

factor_net(): Compile a factorization network
factor_input(): Create an input node for a factorization network
factor_config(): Global configuration for a factorization network
nmf_layer(): Create an NMF factorization layer
svd_layer(): Create an SVD/PCA factorization layer
factor_shared(): Shared factorization across multiple inputs (multi-modal)
factor_concat(): Concatenate H factors from branches (row-bind)
factor_add(): Element-wise H addition (skip/residual connection)
factor_condition(): Concatenate conditioning metadata to a layer's H
fit(): Fit a factorization network
W() H(): Per-factor configuration for factorization layers
cross_validate_graph(): Cross-validate a factorization network
predict(<factor_net_result>): Project new data through a trained factor network
training_logger(): Create a training logger for factor network fitting
export_log(): Export training log to CSV
plot(<training_logger>): Plot training log
as.data.frame(<training_logger>): Convert training log to data.frame
classify_embedding(): Evaluate classification performance of factor embeddings
classify_logistic(): Logistic regression classifier for factor embeddings
classify_rf(): Random forest classifier for factor embeddings
`$`(<factor_net_result>): Access layer results by name
print(<factor_net>): Print a factor_net
print(<factor_net_cv>): Print a factor_net_cv result
print(<factor_net_result>): Print a factor_net_result
print(<fn_factor_config>): Print an fn_factor_config
print(<fn_global_config>): Print an fn_global_config
print(<fn_node>): Print an fn_node
print(<fn_classifier_eval>): Print a classifier evaluation result
summary(<factor_net_result>): Summarize a factor_net_result
summary(<fn_classifier_eval>): Summarize a classifier evaluation result
print(<training_logger>): Print a training log

Simulation

Functions for generating synthetic data

simulateNMF(): Simulate an NMF dataset
simulateSwimmer(): Simulate Swimmer Dataset

StreamPress I/O

Streaming sparse matrix format (.spz)

st_add_transpose(): Add Transpose Section to an Existing StreamPress File
st_chunk_ranges(): Get Column Ranges for Each Chunk in a StreamPress File
st_filter_cols(): Slice Columns Matching Variable Metadata Filter
st_filter_rows(): Slice Rows Matching Observation Metadata Filter
st_free_gpu(): Free GPU-Resident Sparse Matrix
st_info(): Get metadata from a StreamPress file
st_map_chunks(): Apply a Function to Every Chunk in a StreamPress File
st_obs_indices(): Get Row Indices Matching Observation Metadata Filter
st_read(): Read a StreamPress file into a dgCMatrix
st_read_dense(): Read a Dense Matrix from StreamPress v3 Format
st_read_gpu(): Read StreamPress File Directly to GPU Memory
st_read_obs(): Read Observation (Row) Metadata from a StreamPress File
st_read_var(): Read Variable (Column) Metadata from a StreamPress File
st_slice(): Slice Rows and/or Columns from a StreamPress File
st_slice_cols(): Slice Columns from a StreamPress File
st_slice_rows(): Slice Rows from a StreamPress File
st_write(): Write a sparse matrix to a StreamPress file
st_write_dense(): Write a Dense Matrix to StreamPress v3 Format
st_write_list(): Write a List of Matrices as a Single StreamPress File
streampress: StreamPress I/O: Read, Write, and Inspect Compressed Matrices

GPU Support

GPU acceleration functions

gpu_available(): Check if GPU acceleration is available
gpu_info(): Get GPU device information
gpu-backend: GPU NMF Backend
print(<gpu_sparse_matrix>) dim(<gpu_sparse_matrix>) nrow(<gpu_sparse_matrix>) ncol(<gpu_sparse_matrix>): Methods for gpu_sparse_matrix objects

Datasets

Built-in datasets for examples and benchmarks

aml: Acute Myelogenous Leukemia (AML) Dataset
digits: MNIST Digits Dataset
golub: Golub ALL-AML Dataset (Brunet et al. 2004)
hawaiibirds: Hawaii Bird Species Frequency Dataset
movielens: MovieLens Dataset
olivetti: Olivetti Faces Dataset
pbmc3k: PBMC 3k Single-Cell RNA-seq Dataset (StreamPress Compressed)

Package

RcppML RcppML-package: RcppML: Fast Non-Negative Matrix Factorization and Divisive Clustering