Prior work has focused on tabular data

The gap

Evidence profile

Stated in the inline gaps and cells research gap sections of the source papers, classified as general, spanning 2 journals.

Supporting evidence — 3 representative gaps

• Explainable Credit Card Fraud Detection Using LightGBM And SHAP-Guided Feature Selection: A Review (2026) · Zenodo (CERN European Organization for Nuclear Research) · doi

  CONCLUSION AND FUTURE SCOPE This review has argued that predictive performance alone is insufficient for high-stakes financial fraud detection: a deployable system must also be feature interpretable, auditable, efficient, and robust to changing transaction behaviour.

  generalstated in inline gapsevidence 5/5
  Keywords: conclusion future scope review argued predictive performance alone insufficient high stakes financial fraud detection deployable
• Inductive inference of gradient-boosted decision trees on graphs for insurance fraud detection (2026) · Data Mining and Knowledge Discovery · doi

  Prior work has focused on tabular data, but graph-based methods are becoming increasingly popular. - There is a need for methods that can handle heterogeneous and dynamic graphs. - The high class imbalance in fraud data makes it challenging to develop effective detection methods.

  generalstated in cells research gapevidence 5/5
  Keywords: prior work has focused tabular data graph-based methods
• Graph Neural Networks for Financial Fraud and Anomaly Detection (2026) · Zenodo (CERN European Organization for Nuclear Research) · doi

  The paper identifies a gap in existing fraud detection methods, which often miss relational signals in transaction graphs. - The paper notes that traditional detectors, such as gradient boosted trees, treat each transaction as an independent feature vector. - The paper recognizes the need for techniques that can address extreme class imbalance and camouflage.

  generalstated in cells research gapevidence 5/5
  Keywords: paper identifies gap existing fraud detection methods often

Questions about this gap