A bank transaction group discovery method and system based on an overlapping community discovery algorithm

Abstract

The invention discloses a bank transaction group discovery method and system based on an overlapping community discovery algorithm, and the method comprises the steps: 1, reading in bank transaction flow data, and constructing a directed transaction network; 2, initializing the directed transaction network, wherein each node is initialized into a community; 3, for each node, determining whether the node is added into a community to which the adjacent node belongs and a community to which the adjacent node belongs according to the modularity gain, and then reconstructing a transaction network;combining each community in the reconstructed transaction network into one node; skipping to the step 3 to reconstruct the transaction network by applying the modularity gain theory until the community to which all the nodes belong is not changed, or the overlapping modularity of the whole network is not increased or the increasing amplitude is lower than a certain specified threshold value, and finally obtaining the divided community, namely obtaining the bank transaction group. According to the method, overlapped communities in a transaction network can be quickly divided, and bank transaction groups are discovered.

Description

Technical field [0001] The invention belongs to the technical field of financial data processing and analysis, and specifically relates to a method and system for discovering a specific group in bank transactions. Background technique [0002] Commercial banks have a large amount of customer transaction data, which contains many potential laws. Fully mining and using these laws is a necessary means for commercial banks to improve their management level and business competitiveness. With the advent of the big data era, in addition to branch transactions, massive amounts of bank transaction data have been generated on the mobile Internet, which can easily reach the petabyte level. The challenge of gathering, storing, mining, and exploring massive amounts of customer transaction data has increased exponentially, requiring new technology and algorithm support. [0003] The complex network is an effective means to model the transaction behavior of bank customers. The entities in the ne...

AI Technical Summary

Problems solved by technology

Current research focuses on community discovery algorithms to solve the problem of community division in full network data. However, each individual only belongs to a single community, and the result is too ideal. In a complex network, each individual often belongs to multiple communities and has multiple attributes.

Images