exdimred 0.0.3

Toybox for intuitive evaluation of dimensionality reduction

ExDimRed

Examples for Dimensionality Reduction

ExDimRed is a simple visualisation tool, aimed and understanding and evaluating dimensionality reduction methods, using human-intuitive examples.

Examples are 3D point data which are reduced via dimensionality reduction methods to 2D.

Usage

Novel methods, or methods not encluded by default, should follow the standard Scikit-Learn syntax, i.e. have the methods

model.fit_transform(X) to fit to and project input data X.

ExDimRed is accessed as follows:

from sklearn.manifold import Isomap
from exdimred import run

run(Isomap)

This will attempt to generate a pdf of the visualisation in the root directory. If pdf generation fails then a png is generated instead.

Requirements

ExDimRed requires Python >= 3.8 and the packages numpy, matplotlib, scikit-learn, and umap-learn to be installed.