partition review
An open-source R package for interpretable dimensionality reduction through variable partitioning.
Reviewed by iTechGuides Editors · Editorial team · Updated Oct 2026
partition is an open-source R package for researchers who want to reduce dimensionality while retaining interpretable links to their original variables. Its Direct-Measure-Reduce approach creates reduced variables while preserving a user-specified minimum information level. Each original variable maps to exactly one reduced variable, making the resulting structure easier to interpret than an opaque reduction alone. It is a focused tool for R workflows, not a general-purpose statistical-analysis application.
The package centers on agglomerative partitioning, with several choices for how variables are grouped and reduced. Users can configure the variable-selection strategy, information-loss metric and reduction function. Built-in partitioners include intraclass correlation (ICC), k-means, minimum-R², principal-component and standardized-mutual-information approaches. Visualization and permutation-test functions support examining partitions and their results. This combination suits analyses where the partitioning choices and information trade-offs need to remain explicit; researchers seeking a broader set of statistical procedures will need other software alongside it.
partition is distributed as open-source software through CRAN and GitHub, with desktop availability on Windows, macOS and Linux. It works with R data frames and tibbles, so it fits best where R is already part of the analysis workflow. Its open-source model avoids a commercial plan decision, while documentation and community are the listed support channels. Choose partition when interpretable variable partitioning is the main need and customizable reduction matters. Choose a broader statistical package or hosted application if you need analysis beyond this narrow dimensionality-reduction focus or do not want an R-based workflow.
partition pros and cons
- Where it wins
- Maps each original variable to exactly one reduced variable.
- Configures variable selection, information loss and reduction methods.
- Includes visualization and permutation-test functions.
- Where it doesn't
- Focused on dimensionality reduction rather than general statistical analysis.
- Requires working with R data frames or tibbles.
- As an R package, it is not a hosted statistical-analysis application.
partition fact sheet, pricing and score →
Advertiser disclosure: iTechGuides is reader-supported. Vendors can pay for top positions in our rankings and for a place on other products' pages, and we may earn a commission when you click some links. Bought places stand at the number bought; this page and our disclosure say so. Paying vendors may get a fuller, more detailed review from our editors, and that review can change their score. How we rank.
Last updated · How we research and update