Here is Matlab code written by the authors of each method:

- Laplacian Eigenmaps by Mikhail Belkin.
- Locality-Preserving Projections by Xiaofei He
- Locality Pursuit Embedding by Wanli Min

Dinoj Surendran has begun rewriting and/or wrapping, and optimizing where possible, each algorithm so it can be called in the common form

E = dimredmethod (X,L);

Where X is a N x H matrix representing N points in R^{H}, E
is a N x L matrix representing the same N points in R^{L}.
Of course, each method can have other options too. E(:,i) is the i-th
component, as you would expect. The code is not necessarily the same
as that of the original authors, so you use it at your own
risk. To run this code, you will also need the function L2_distance.m
by Roland Bunschoten.

So far DS has tested code for the following algorithms

- lapbin.m : Laplacian Eigenmaps with Binary Weights (testing notes)
- lapheat.m : Laplacian Eigenmaps with Heat-Kernel Weights (not the normalized Laplacian)