32 control points	64 control points	256 control points	512 control points

pyLSP

pyLSP is a lightweight python implementation of the Least Square Projection (LSP) dimensionality reduction technique using sklearn style API.

The implementation is based on the paper "Least Square Projection: A Fast High-Precision Multidimensional Projection Technique and Its Application to Document Mapping", which can be cited using:

@ARTICLE{4378370,
  author={Paulovich, Fernando V. and Nonato, Luis G. and Minghim, Rosane and Levkowitz, Haim},
  journal={IEEE Transactions on Visualization and Computer Graphics}, 
  title={Least Square Projection: A Fast High-Precision Multidimensional Projection Technique and Its Application to Document Mapping}, 
  year={2008},
  volume={14},
  number={3},
  pages={564-575},
  keywords={Least squares methods;Multidimensional systems;Data visualization;Least squares approximation;Data analysis;Computational geometry;Testing;Text processing;Data mining;Demography;Multivariate visualization;Data and knowledge visualization;Information visualization;Multivariate visualization;Data and knowledge visualization;Information visualization},
  doi={10.1109/TVCG.2007.70443}}

A small working example can be found in tests/iris_example.py and tests/digits_example.py.