Hakan Löfqvist, Neo4j

1. AML with Graph Data Science
Håkan Löfqvist, Neo4j
hakan.lofqvist@neo4j.com

2. “Data science is an interdisciplinary
field that uses scientific methods,
processes, algorithms and systems
to extract knowledge and insights
from structured and unstructured
data.” - Wikipedia
2
Data scientists use data to
answer questions.

3. Graph Data Science is a
science-driven approach to gain
knowledge from the relationships
and structures in data, typically to
power predictions.
3
Data scientists use
relationships to answer
questions.

4. 4
Fraud
Marketing
Customer
Journey
in Financial Services and Banking
• First party & synthetic
identity fraud
• Fraud rings
• Anti money laundering
• Disambiguation
• Recommendations
• Customer segmentation
• Churn prediction

5. 5
Graph Feature
Engineering
Graph
Embeddings
Graph
Networks
Knowledge
Graphs
Graph
Analytics

6. Graph Feature
Engineering
Graph
Embeddings
Graph
Networks
6
Graph
AnalyticsKnowledge
Graphs
Graph search
and queries
Support domain
experts

7. Deceptively Simple Queries
How many flagged accounts are in the
applicant’s network 4+ hops out?
How many login / account variables in
common?
Add these metrics to your approval
process
Queries in Financ
Improving existing pipelines to identify fraud via heuristics
7

8. Graph Feature
Engineering
Graph
Embeddings
Graph
Networks
8
Knowledge
Graphs
Graph
Analytics
Graph queries &
algorithms for
offline analysis
Understanding
Structures

9. • Degree Centrality
• Closeness Centrality
• Harmonic Centrality
• Betweenness Centrality & Approx.
• PageRank
• Personalized PageRank
• ArticleRank
• Eigenvector Centrality
• Triangle Count
• Clustering Coefficients
• Connected Components (Union Find)
• Strongly Connected Components
• Label Propagation
• Louvain Modularity
• Balanced Triad (identification)
• Shortest Path
• Single-Source Shortest Path
• All Pairs Shortest Path
• A* Shortest Path
• Yen’s K Shortest Path
• Minimum Weight Spanning Tree
• K-Spanning Tree (MST)
• Random Walk
• Breadth & Depth First Search
• Triangle Count
• Local Clustering Coefficient
• Connected Components (Union Find)
• Strongly Connected Components
• Label Propagation
• Louvain Modularity
• K-1 Coloring
• Modularity Optimization
• Euclidean Distance
• Cosine Similarity
• Node Similarity (Jaccard)
• Overlap Similarity
• Pearson Similarity
• Approximate KNN
Pathfinding
& Search
Centrality /
Importance
Community
Detection
Similarity
Link
Prediction
• Adamic Adar
• Common Neighbors
• Preferential Attachment
• Resource Allocations
• Same Community
• Total Neighbors
Embeddings
• Node2Vec
• FastRP
• GraphSAGE

10. Graph Algorithms for Detecting Fraud
Graph algorithms enable reasoning
about network structure
Louvain to identify communities
that frequently interact
PageRank to measure influence
and transaction volumes
Connected components
identify disjointed group
sharing identifiers
Jaccard to measure account
similarity
10

11. Graph
Embeddings
Graph
Networks
11
Knowledge
Graphs
Graph
Analytics
Graph Feature
Engineering
Graph algorithms
& queries for
machine learning
Improve Prediction
Accuracy

12. Feature Engineering is how we combine and process the data
to create new, more meaningful features. Using graphs we
can base ML on influential people that share identifiers.
12
Financial Transaction Data

13. Graph Feature
Engineering
13
Knowledge
Graphs
Graph
Analytics
Graph
Embeddings
Graph embeddings
for dimensionality
reduction
Predictions on
complex structures
Graph
Networks

14. 14
Graph Embeddings

15. 15
Demo

16. 1
6
Graph Model

17. 17
What we will cover in the demo
Knowledge Graph Queries
Graph Analytics
Graph Embedding + ML