Spycsh / spam-detection-neo4j

GCN-based Anti-Spam (GAS) model with Neo4j prototype

spam-detection-neo4j

How to run

1. Install Neo4j.
2. Set the name and password in Neo4j to name: "neo4j", and password: "admin".
3. Run the following command in Neo4j to create the initial graph:

CREATE (user0:User{name:'user0',id:0})
CREATE (user1:User{name:'user1',id:1})
CREATE (user2:User{name:'user2',id:2})
CREATE (user3:User{name:'user3',id:3})
CREATE (user4:User{name:'user4',id:4})

CREATE (item0:Item {name:'item0',id:0})
CREATE (item1:Item {name:'item1',id:1})
CREATE (item2:Item {name:'item2',id:2})


CREATE (user0)-[review0:Review{name:'review0',id:0}]->(item0)
CREATE (user0)-[review1:Review{name:'review1',id:1}]->(item1)
CREATE (user1)-[review2:Review{name:'review2',id:2}]->(item0)
CREATE (user2)-[review3:Review{name:'review3',id:3}]->(item0)
CREATE (user3)-[review5:Review{name:'review5',id:5}]->(item2)
CREATE (user4)-[review4:Review{name:'review4',id:4}]->(item1)
CREATE (user4)-[review6:Review{name:'review6',id:6}]->(item2)

You can check the graph in Neo4j by

MATCH (n) RETURN n

4. Run GAS_main.py

Some notes

click the node labels or

MATCH (n) RETURN n LIMIT 25

MATCH (user)-[:Review]->(item)
RETURN user.name, item.name

MATCH (user)-[review0:Review{name:'review0'}]->(item)
RETURN user.name, item.name

delete all

match (a) -[r] -> () delete a, r
match (a) delete a