DenseFlow Algorithm Detects Suspicious Subgraphs and Traces Money Laundering Pathways

============================================================

A team of researchers has developed a novel algorithm, DenseFlow, that can effectively detect suspicious subgraphs in large directed graphs and trace the money laundering pathways connecting them. The algorithm is designed to identify and remove nodes with minimum suspiciousness, reducing computational overhead while maximizing objective function.

Dense Subgraph Detection

The algorithm begins by defining the joint suspiciousness of a node to a subset, f(S), and a hyperparameter, b. This allows for dynamic updates of node suspiciousness as the suspicious subset evolves. The total suspiciousness (g) of a subset S is calculated as the sum of individual node suspiciousness values.

Greedy Algorithm

DenseFlow employs a greedy approximation algorithm inspired by Charikar’s method to find the optimal subset S that maximizes g. The algorithm iteratively identifies and removes nodes with minimum suspiciousness and updates related nodes’ suspiciousness values. To reduce computational overhead, a priority tree is constructed for nodes in subset S, allowing efficient location of the leaf node corresponding to the global minimum value.

Priority Tree Construction

Each internal tree node records the minimum value among its child nodes, reducing time complexity from O(|S|) to O(log|S|). The weight assigned to each node i is defined as fi(S).

Tracing Money Laundering Pathways

After detecting suspicious subgraphs, DenseFlow aims to identify the money laundering pathways connecting these subgraphs. In a directed transaction graph, nodes represent accounts, and edge weights indicate transaction amounts. The algorithm uses the maximum flow algorithm to discover the money laundering pathways from a source account to downstream money laundering gangs.

Maximum Flow Algorithm

The maximum flow problem is used to find the maximum possible flow from the source node to junctions in a directed graph. This allows simulation of the flow of funds in a network, enabling the detection of money laundering pathways.

Conclusion

==========

DenseFlow’s novel approach to detecting suspicious subgraphs and tracing money laundering pathways has significant implications for financial fraud detection and prevention. The algorithm’s efficiency and effectiveness make it a valuable tool for law enforcement agencies and financial institutions seeking to combat money laundering and terrorist financing.