Description
Likelihood-Based Intrinsic Dimension Estimators.
Description
Provides functions to estimate the intrinsic dimension of a dataset via likelihood-based approaches. Specifically, the package implements the 'TWO-NN' and 'Gride' estimators and the 'Hidalgo' Bayesian mixture model. In addition, the first reference contains an extended vignette on the usage of the 'TWO-NN' and 'Hidalgo' models. References: Denti (2023, <doi:10.18637/jss.v106.i09>); Allegra et al. (2020, <doi:10.1038/s41598-020-72222-0>); Denti et al. (2022, <doi:10.1038/s41598-022-20991-1>); Facco et al. (2017, <doi:10.1038/s41598-017-11873-y>); Santos-Fernandez et al. (2021, <doi:10.1038/s41598-022-20991-1>).
README.md
intRinsic v1.1.0 
A package with functions to estimate the intrinsic dimension of a dataset via likelihood-based approaches. Specifically, the package implements the TWO-NN and Gride estimators and the Hidalgo Bayesian mixture model.
To install the package from CRAN, run
install.packages("intRinsic")
To install the package from this GitHub repository, run
# install.packages("remotes")
#Turn off warning-error-conversion regarding package versions
Sys.setenv("R_REMOTES_NO_ERRORS_FROM_WARNINGS" = "true")
#install from github
remotes::install_github("Fradenti/intRinsic")
Simple example on Swissroll dataset
library(intRinsic)
X <- Swissroll(2000)
twonn(X)
The vignette for this package has been published in the Journal of Statistical Software. The article can be found at this link.
Please help me improve this package by reporting suggestions, typos, and issues at this link.
Please note that the previous versions of the package (from v0.1.0 to v1.0.2) are still available as GitHub Releases at this page.