DRL-based approach integrating network embedding to address the competitive influence maximization on evolving social networks
This work is based on existing GitHub Code (https://github.com/devsisters/DQN-tensorflow).
Python Tensorflow implementation of Network Embedding Meets Deep Reinforcement Learning to Tackle Competitive Influence Maximization on Evolving Social Networks.
This implementation contains:
First, install prerequisites with:
$ pip install tqdm gym[all]
Second, generate embeddings of a social graph (evoling social networks) using generate_embeddings.py inside data folder
To train a model for an evolving social networks such as bitcoinalpha and bitcoinotc against opponent's degree strategy:
$ python main.py --env_name=bitcoinalpha --is_train=True --opponent degre
$ python main.py --env_name=bitcoinotc--is_train=True --opponent degre
To test a model for an evolving social networks such as bitcoinalpha against opponent's weight strategy:
$ python main.py --is_train=False --env_name=bitcoinalph --opponent weight --testing_episode 2000