Intrinsic dimensionality estimation techniques

Implementation of some state-of-art intrinsic dimensionality estimators.

You are now following this Submission

Data analysis is a fundamental step to face real Machine-Learning problems, various well-known ML techniques, such as those related to clustering or dimensionality reduction, require the intrinsic dimensionality (id) of the dataset as a parameter.

To the aim of automate the estimation of the id, in literature various techniques has been described, this small toolbox contains the implementation of some state-of-art of them, that is: MLE, MiND_ML, MiND_KL, DANCo, DANCoFit.

For an R implementation see:
http://www.maths.lth.se/matematiklth/personal/johnsson/dimest/

Cite As

Gabriele Lombardi (2026). Intrinsic dimensionality estimation techniques (https://se.mathworks.com/matlabcentral/fileexchange/40112-intrinsic-dimensionality-estimation-techniques), MATLAB Central File Exchange. Retrieved .

Acknowledgements

Inspired: Rand Sphere.zip

General Information

MATLAB Release Compatibility

  • Compatible with any release

Platform Compatibility

  • Windows
  • macOS
  • Linux
Version Published Release Notes Action
1.1.0.0

Added a reference to an R implementation in the description.

1.0.0.0