Learn R Programming

repsim (version 0.1.0)

Measures of Representational Similarity Across Models

Description

Provides a collection of methods for quantifying representational similarity between learned features or multivariate data. The package offers an efficient 'C++' backend, designed for applications in machine learning, computational neuroscience, and multivariate statistics. See Klabunde et al. (2025) for a comprehensive overview of the topic.

Copy Link

Version

Install

install.packages('repsim')

Monthly Downloads

114

Version

0.1.0

License

MIT + file LICENSE

Maintainer

Kisung You

Last Published

November 7th, 2025

Functions in repsim (0.1.0)

repsim_hsic

List of HSIC estimators
svcca

Singular Vector Canonical Correlation Analysis
repsim_kernels

List of kernel functions
cka

Centered Kernel Alignment
dot_product

Dot product similarity
pwcca

Projection-Weighted Canonical Correlation Analysis
cca

Canonical Correlation Analysis
hsic

Hilbert Schmidt Independence Criterion
lin_reg

Linear regression fit similarity