mlsanityData to 7 Oct 2026

Interpretable vs Learned Encoders for High-Cardinality Fraud Detection

arXiv 2607.00477 ↗First posted 1 Jul 2026cs.LG cs.CEcited by 2

Not on the map. Why?

Abstract (arXiv)

A total of seven categorical encoding methods were tested on the IEEE-CIS fraud benchmark dataset (590,540 records, 3.5% positives, 8 high-cardinality columns). The encoders were evaluated using a stratified 5-fold cross-validation (CV) with three repetitions. Five of the encoders had identical frozen LightGBM learners in the downstream phase, allowing for controlled comparisons of their performance to each other. CatBoost and TabNet were included as comparisons across paradigms using different learners. The entity embeddings produced the highest AUC-ROC (0.9612), with a statistically significant tie with that of CatBoost (0.9602) and statistically superior to tier group encoding (0.9548), whereas target encoding was only 0.0023 worse than tier group encoding and the auditor-friendly tier boundaries were maintained. Off-the-shelf TabNet did not outperform tree-based pipelines and collapsed under data scarcity. On AUC-PR, CatBoost leads (0.822 vs. 0.793); no encoder dominated both metrics. Per-column analysis confirmed the embedding advantage arises from joint multi-column representation.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 14 of its 30 references that are in the map. First posted on arXiv 1 Jul 2026.

Builds on

14 of 30 references.

References · 14

  1. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  2. 2021Revisiting Deep Learning Models for Tabular Datacited by 725
  3. 2019TabNet: Attentive Interpretable Tabular Learningcited by 640
  4. 2021Tabular Data: Deep Learning is Not All You Needcited by 530
  5. 2021Deep Neural Networks and Tabular Data: A Surveycited by 489
  6. 2019Neural Oblivious Decision Ensembles for Deep Learning on Tabular Datacited by 200
  7. 2016Entity Embeddings of Categorical Variablescited by 85
  8. 2018Similarity encoding for learning with dirty categorical variablescited by 28
  9. 2019Encoding high-cardinality string categorical variablescited by 22
  10. 2021Regularized target encoding outperforms traditional methods in supervised machine learning with high cardinality featurescited by 20
  11. 2026Task-Specific Efficiency Analysis: When Small Language Models Outperform Large Language Modelscited by 15
  12. 2026Sustainable Hybrid Document-Routed Retrieval for Financial RAG: Resolving the Robustness-Precision Trade-offcited by 13
2 more
  1. 2020Deep Learning Methods for Credit Card Fraud Detectioncited by 6
  2. 2019Encoding Categorical Variables with Conjugate Bayesian Models for WeWork Lead Scoring Enginecited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026From Non-Convex Self-Concordant Regularization to Scalable Quasi-Newton Training of PINNscited by 0
  2. 2026Explicit State Elicitation Is Not Enough: A Controlled Audit of Memory-Policy Classificationcited by 0