Python Implementation of Max-Margin DeepWalk :- https://github.com/thunlp/MMDW
Paper :- "Max-Margin DeepWalk: Discriminative Learning of Network Representation" - IJCAI, 2016
Input is organized as follows -
Datasets/
|_ _ _ <Dataset-name>
|_ _ _ <Dataset-name.mat>
|_ _ _ <Stats.txt>
|_ _ _ <Percentage of train-test split>
|_ _ _ <Fold-No>
|_ _ _ test_ids.npy
|_ _ _ train_ids.npy
|_ _ _ val_ids.npy
python main_algo.py --DATA_DIR cora --ALPHA_BIAS -2 --ALPHA 1.0 --LAMBDA 1.0 --L_COMPONENTS 16
* ALPHA_BIAS : Alpha bias level for biased random walk
* ALPHA : Similarity matrix factorization weight
* LAMBDA : L2 regularization weight
* L_COMPONENTS : Dimension of projected space
* The generated node and label embeddings are saved in Emb/ folder as <emb_dataset_U/Q<Fold-No>>.npy.
* The node and label embeddings are of dimension (#Nodes x L_COMPONENTS) & (#Labels x L_COMPONENTS).
* The Node Classification evaluation results are stored in - Results/ folder