Please run 'python HIME.py' to generate the embedding model. The parameters are as follows: --dataset: 'dblp', 'protein_go' or 'gene_pathway', default 'dblp'. --neg_num negative sampling number, default 5. --emb_num branch vector number, default 8. --emb_dim embedding dimension, default 32. --epoch_num epoch number, default 50. --batch_size batch size, default 1000. --LRU_period LRU period, default 5. The trained model will be saved into a directory called 'saved_model'.
To evaluate, please run 'python evaluate.py'. The parameters are as follows: --dataset: 'dblp', 'protein_go' or 'gene_pathway', default 'dblp'. --emb_num branch vector number, default 8. --emb_dim embedding dimension, default 32. --epoch_num epoch number, default 50. The program will find the model specified by above parameters in 'saved_model', and the perform the evaluation.