Disrupt fraud rings with social network analysis
Social network analysis is vital to identifying the key, highly connected individuals participating in fraud, tracking down the influencers and ringleaders, and disrupting these networks.
Combined with visualization, your analysts have an arsenal at hand to see the steps that individuals took to commit fraud, see the relationships between fraudulent activities, view the pathways through which transactions occur, and review the patterns in historical fraudulent behavior to inform predictive analysis and prevention tactics.
30+ analysis algorithms at your fingertips
Perspectives includes over 30 built-in algorithms for data-driven analysis. Some algorithms perform traversals and path analysis, detect or break cycles, and find dependencies and reachable graph elements. Others cluster or partition graph elements, conduct root cause analysis, perform network flow computations, or compute centrality measures in social networks. Below we highlight a few algorithms most useful for identifying potential fraud.
-
Betweenness Centrality
-
Bridge Detection
-
Closeness Centrality
-
Clustering Algorithm
-
Eigenvector Centrality
-
Shortest Path Detection
Betweenness Centrality
Measures a node based on how many shortest paths that the node is a member of. Betweenness centrality reveals the intermediaries, or middlemen, from the data.
Bridge Detection
Finds all bridges in an undirected graph. An edge is considered a bridge if it is the only path between its end nodes. When a bridge is removed from an undirected graph, its end nodes are allocated to different connected components of the graph.
Closeness Centrality
Measures how many steps a message would need to travel to reach all the other nodes in the network.
Clustering Algorithm
Groups nodes that share similar properties into sets called clusters. A cluster often represents parts of the graph topology that belong together, and therefore can help investigators focus their efforts.
Eigenvector Centrality
Determines the importance of each node based on the importance of its connections. Fraud investigators can focus their investigation on those transactions with a high level of influence within a network.
Shortest Path Detection
Calculates the path that uses the fewest edges between a specified start node and finish node. If the starting node is fraudulent, it can aid the investigation into nearby nodes that may also be fraud or targets of fraud.
