Closed alrod97 closed 4 years ago
Right now we've already implemented PCA, TSNE, KernelPCA(currently doing this), and LLE.
There are other algorithms that seem to be interesting as autoencoders. An overview is given in
http://research.iac.es/winterschool/2018/media/slides/baron_dimensionality_reduction.pdf
Maybe we can implement some of them and see if they do better then the current ones.
" projection to 2d coordinate subspaces, PCA, t-SNE, UMAP, Isomap, LLE, NCA, kepler mapper, AEs,"
ISOMAP and UMAP added
Final Scope: PCA, LLE, TSNE, UMAP, ISOMAP, Keppler Mapper, MDA
Right now we've already implemented PCA, TSNE, KernelPCA(currently doing this), and LLE.
There are other algorithms that seem to be interesting as autoencoders. An overview is given in
http://research.iac.es/winterschool/2018/media/slides/baron_dimensionality_reduction.pdf
Maybe we can implement some of them and see if they do better then the current ones.