Simplivariate Component Analysis for detecting constant, additive, multiplicative, and user-defined biclusters in numeric data matrices. The method uses a genetic algorithm to simultaneously identify multiple simplivariate components, with flexible pattern definitions and built-in visualization. SIMPLICA is a versatile biclustering framework for numeric data matrices. Whether your data come from biology, chemistry, or any other field, SIMPLICA helps uncover hidden structures using genetic algorithms and flexible pattern definitions.

Before SIMPLICA:

After SIMPLICA:

Installation

You can install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("joshageman/SIMPLICA")

Once on CRAN:

install.packages("SIMPLICA")

Example

This example uses the built-in simplicaToy dataset.

library(SIMPLICA)
# Load toy dataset
data("simplicaToy")
# Run SIMPLICA
fit <- simplica(
  df = simplicaToy$data,
  verbose = TRUE
)
# Print summary of found components
summary(fit)
# Plot detected clusters
plotComponentResult(df = simplicaToy$data,
                    string            = fit$best$string,
                    componentPatterns = fit$best$componentPatternsUpdated,
                    componentScores   = fit$best$componentScores,
                    showAxisLabels    = FALSE,
                    title             = "SIMPLICA on simplicaToy",
                    scoreCutoff       = 25000

Features

• Detects constant, additive, multiplicative, and user-defined patterns.
• Simultaneous optimization of multiple biclusters via genetic algorithms.
• Flexible penalty settings to control overfitting.
• Customizable pattern functions.
• Publication-ready visualizations.
• Suitable for small to medium-sized datasets typical in life sciences and chemometrics.

Citation

If you use SIMPLICA in your research, please cite:

Hageman, J. (2025). SIMPLICA: Simplivariate Component Analysis for Biclustering. R package version 1.0.0.

License

GPL-3 © Jos Hageman