A new manifold learning algorithm

[ghost]

This code presents a geometric approach for data embedding called Polygonal Coordinate System (PCS), which is able to efficiently represent multidimensional data into a 2D plane by preserving the global structure of the data. For this purpose, data are represented across a regular polygon or interface between the high dimensionality and the two-dimensional data. PCS can properly handle massive amounts of data (Big Data) by adopting an incremental and linear-time complexity dimensionality reduction.

Reference Issues/PRs

What does this implement/fix? Explain your changes.

Any other comments?

[rth]

Thanks for your PR. Please add a citation for this algorithm, and check that it would pass the inclusion criteria at https://scikit-learn.org/stable/faq.html#what-are-the-inclusion-criteria-for-new-algorithms

[amueller]

@CaioMFRodrigues are you still interested in this?
I couldn't find any reference with a quick search. I'm in favor of closing this if it doesn't show any google hits.

[ghost]

@CaioMFRodrigues are you still interested in this?
I couldn't find any reference with a quick search. I'm in favor of closing this if it doesn't show any google hits.

This work has been recently published at the BRACIS conference (soon it will be available on the IEEE website). I didn't know all the rules found at <https://scikit-learn.org/stable/faq.html#what-are-the-inclusion-criteria-for-new-algorithms>. If you want to close this, that's fine.

[ghost]

I'm sorry.

[amueller]

@CaioMFRodrigues all good. You can contribute it as a new project to scikit-learn-contrib or as a new model to scikit-learn-extra.