Graph Sample Boundaries & Limitations
Detailed registry of graph boundaries, depth traversal cutoffs, and methodological data limits. Built to prevent forensic analysts from misclassifying peripheral boundary artifacts as terminal cash sinks.
Operational Warning: Depth-4 Boundary Truncation
Depth-4 nodes with zero observed downstream activity are strictly classified as graph boundary nodes, not automatically as terminal recipients or laundering endpoints. Outgoing chains may extend beyond the four-hop query perimeter. Manual expansion or multi-source querying is required before making legal determinations.
Algorithmic cutoff artifact: these nodes are boundary entities, not confirmed terminal recipients or cash drop points.
Funds deposited prior to sampling window or transferred via internal ledger channels outside reporting scope.
Potential micro-smurfing vulnerability: fragmentation below 5K KZT remains invisible in graph traversal.
Inter-bank clearing gateways and cash-in terminals lacking KYC linkage in current pipeline slice.
Automated Verification Rules Against false Classification
To uphold legal-grade investigative rigor, our pipeline automatically tags truncated endpoints and isolates dormant seed entities before generating investigation dossiers.
Depth-4 Terminal Reclassification
Automatic marking of leaf nodes at depth=4 as 'Boundary' rather than 'Terminal Sink' prevents false-positive seizure requests.
Seed Target Reconciliation
Cross-checks stagnant seed accounts with core ledger records to separate inactive targets from data pipeline drops.
Threshold Leakage Auditing
Periodic sampling of sub-5K transactions to verify that aggregated smurfing networks do not circumvent graph construction.
Filters & Thresholds
NEXUS AML applies rigorous data filtering and structural thresholds to ensure high-fidelity network reconstruction and actionable intelligence.
Automated exclusion of low-value transactions below 5k KZT to reduce noise and focus on significant capital movements.
Detection of rapid pass-through funds where 87% of inflow is forwarded within 48 hours, identifying transit nodes.
Boundary node identification at depth-4 to prevent false terminal classification of incomplete network segments.
Automated scoring pipeline that flags nodes based on betweenness, seed proximity, and structural importance metrics.
Need custom data processing thresholds?
Configure your AML pipeline parameters for specific network topologies.