This project is for financial risk control model construction based on KG. In this project, two goal are need to implement. The first is construct a financial prediction model, the sencod is construct real-time system.
The picture show the whole architect of the system.
The project implemented on Spring Boot
Rule Developement
To extract the feature for applicant, we pre-defined three rules to judge applicant are more likely to be delinquent
- The number of delinquent application of applicant.
For its
Cypher query, are
match (p:Person)-[h:HAS_APPLICATION]->(a:Application) where a.status=”OVER_DUE” and a.personId=$personId return count(a)
- How many People are in blacklist in the one degree relationship of applicant
For its
Cypher query, are
match (p:Person)-[]-(p1:Person)-[h:HAS_PHONE]-(b:Black) where p.personId= $personId return count(b)
- How many People are in blacklist in the two degrees relationship of applicant
For its
Cypher query, are
macth (p:Person)-[]-(p2:Person)-[]-(p2:Person)-[h:HAS_PHONE]-(b:Black) where p.personId = $personId return count(b)
Save three rules in Mysql
Intergration of Mysql and Neo4j in Spring Boot
Extract rule from Mysql and then use this query to query neo4j database that obtain three features for applicant.