nuggets

nuggets is a package for R statistical computing environment providing a framework for systematic exploration of association rules (Agrawal (1994)), contrast patterns (Chen (2022)), emerging patterns (Dong (1999)), subgroup discovery (Atzmueller (2015)), and conditional correlations (Hájek (1978)). User-defined functions may also be supplied to guide custom pattern searches.

Supports both crisp (Boolean) and fuzzy data. Generates candidate conditions expressed as elementary conjunctions, evaluates them on a dataset, and inspects the induced sub-data for statistical, logical, or structural properties such as associations, correlations, or contrasts. Includes methods for visualization of logical structures and supports interactive exploration through integrated Shiny applications.

Key Features

• Support for both categorical and numeric data.
• Provides both Boolean and fuzzy logic approach.
• Data preparation functions for easy pre-processing phase.
• Functions for examining associations, conditional correlations, and contrasts among data variables.
• Visualization and pattern post-processing tools.
• Integrated Shiny applications for interactive exploration of discovered patterns.

Documentation

Read the full documentation of the nuggets package.

Installation

To install the stable version of nuggets from CRAN, type the following command within the R session:

install.packages("nuggets")

You can also install the development version of nuggets from GitHub with:

install.packages("devtools")
devtools::install_github("beerda/nuggets")

To start using the package, load it to the R session with:

library(nuggets)

Minimal Example

The following example demonstrates how to use nuggets to find association rules in the built-in mtcars dataset:

# Preprocess: dichotomize and fuzzify numeric variables
cars <- mtcars |>
    partition(cyl, vs:gear, .method = "dummy") |>
    partition(carb, .method = "crisp", .breaks = c(0, 3, 10)) |>
    partition(mpg, disp:qsec, .method = "triangle", .breaks = 3)

# Search for associations among conditions
rules <- dig_associations(cars,
                          antecedent = everything(),
                          consequent = everything(),
                          max_length = 4,
                          min_support = 0.1,
                          measures = c("lift", "conviction"))

# Explore the found rules interactively
explore(rules, cars)

Contributing

Contributions, suggestions, and bug reports are welcome. Please submit issues on GitHub.