Business Context
A financial institution wants to detect potential money laundering rings. They need to identify 'Hub' accounts (high transaction volume/degree) and 'Bridge' accounts (high betweenness) that connect otherwise unrelated groups of customers, which is a common pattern in layering funds.
Data Preparation
Creating a transaction graph representing fund transfers between accounts.
Copied!
| 1 | |
| 2 | DATA mycas.transactions; |
| 3 | INPUT from_acc $ to_acc $ amount; |
| 4 | DATALINES; |
| 5 | Acc1 Acc2 1000 Acc1 Acc3 500 Acc2 Acc4 2000 Acc3 Acc4 150 Acc4 Acc5 3000 Acc5 Acc6 400 Acc6 Acc7 100 Acc5 Acc8 2000 ; |
| 6 | |
| 7 | RUN; |
| 8 | |
| 9 | DATA mycas.accounts; |
| 10 | INPUT node $; |
| 11 | DATALINES; |
| 12 | Acc1 Acc2 Acc3 Acc4 Acc5 Acc6 Acc7 Acc8 ; |
| 13 | |
| 14 | RUN; |
| 15 |
Étapes de réalisation
1
Load data and prepare the graph structure.
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | |
| 4 | TABLE.loadTable RESULT=r STATUS=s / caslib="casuser" path="transactions.sashdat"; |
| 5 | |
| 6 | RUN; |
| 7 |
2
Calculate Weighted Degree and Betweenness to find volume and bridges.
Copied!
| 1 | PROC CAS; |
| 2 | network.centrality / |
| 3 | links={name='transactions'} |
| 4 | nodes={name='accounts'} |
| 5 | degree='BOTH' |
| 6 | between='WEIGHT' |
| 7 | outNodes={name='aml_scores', replace=true}; |
| 8 | RUN; |
Expected Result
The 'aml_scores' table should be generated containing 'cent_degree_out' (money sent), 'cent_degree_in' (money received), and 'cent_between_wt'. Accounts like Acc4 and Acc5 should show high betweenness centrality, indicating they act as bridges in the flow of funds.