This package has been made to compare some graph embedding algorithms. The following methods are implemented: Laplacian EigenMaps, Locally Linear Embedding, Higher-Order Proximity preserved Embedding (HOPE), Multi-dimensional scaling of a dissimilarity matrix, Node2vec, Struc2vec, Verse, Singular Value Decomposition of the Adjacency matrix, Kamada-Kawai Layout (KKL), Structural Deep Network Embedding (SDNE)
Requirements:
Python3: click, numpy, scipy, networkx, collections, sklearn, pandas, keras, tensorflow, python-igraph, graph_tool, python-louvain
Python2: futures, fastdtw, gensim
java: version 8 update 201
Snap (node2vec): download and compile Snap. Put node2vec executable in the corresponding directory and give it the right to execute (chmod +x)
Verse: download and compile Verse
Source:
Snap: https://github.com/snap-stanford/snap Verse: https://github.com/xgfs/verse LE, LLE, HOPE, SDNE: https://github.com/palash1992/GEM struc2vec: https://github.com/leoribeiro/struc2vec