Knowledge Graph Construction for Payment Data Risk Control

Abstract

In the process of controlling the fraud risk in the Internet finance field, in order to solve the problem of large amount of data, complex and confusing information provided by data, an end-to-end risk control knowledge graph based on payment data construction method was adopted, and the visualization of user-oriented was realized. This method sets the labels of the data according to the requirements of the business scenario and then generates the format required by the database and saves the data as a triple with extended attributes. The background system connects to the graph database Neo4j and automatically imports triples of data in the way of incremental data extraction and converts them into the knowledge graph system, so as to provide users with Internet financial risk control services based on knowledge graph in the Web. The risk control knowledge graph constructed by this method can import the data into the database dynamically, greatly reducing the construction time of the graph, and displaying the business relationship of the users in the way of combining text list and graphical visualization, making it more convenient for users to judge the risk type. It can be seen from the case mined from the knowledge graph system that the risk control knowledge graph has a good effect.