EUVIMEDEUVIMED entdecken
Crossref

Hybrid LSTM and Ensemble Model for Credit Card Fraud Detection

Ruchita Saraf Ruchita Saraf, Shravya Sanikere Shravya Sanikere, Tejas N G Tejas N G, Shivam Shivam, Savitha G Savitha G

International Journal of Creative and Open Research in Engineering and Management · 17.05.2026 · 02 (05) · 1-5

DOI: 10.55041/ijcope.v2i5.537 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Creative and Open Research in Engineering and Management, online erschienen am 17.05.2026. Verfasst von Ruchita Saraf Ruchita Saraf und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Credit card fraud detection remains one of the most challenging tasks in financial machine learning due to extreme class imbalance, often less than 0.2% fraud, and the dynamic sequential nature of legitimate and fraudulent transactions. This study proposes a hybrid model that combines a Long Short-Term Memory (LSTM) network with an attention mechanism and a calibrated Random Forest (RF) classifier. The LSTM captures long-term temporal dependencies in transaction sequences with window size T = 5, while the attention layer dynamically weighs the importance of each time step. Concurrently, the Random Forest processes static features of the most recent transaction and outputs a calibrated probability. A weighted ensemble, Pfinal = 0.6PLSTM + 0.4PRF, fuses the two predictions. Using the publicly available Kaggle credit card dataset with 284,807 transactions and 0.18% fraud, the proposed approach achieves a Precision-Recall AUC of 0.806, a fraud class F1-score of 0.85, and near-perfect precision of 0.98 for fraud cases at the optimal decision threshold of 0.727. Extensive experiments compare the proposed method against baseline models including standalone LSTM, LSTM with attention, and Random Forest. All data statistics, score ranges, and classification metrics are reported exactly as obtained from the simulation environment. Index Terms—Credit card fraud detection, attention mecha-nism, ensemble methods, imbalanced datasets, LSTM networks, Random Forest classifier.

Zitieren

APA

Saraf, R. S. R., Sanikere, S. S. S., G, T. N. G. T. N., Shivam, S., & G, S. G. S. (2026). Hybrid LSTM and Ensemble Model for Credit Card Fraud Detection. International Journal of Creative and Open Research in Engineering and Management, 02(05), 1-5. https://doi.org/10.55041/ijcope.v2i5.537

Vancouver

Saraf RSR, Sanikere SSS, G TNGTN, Shivam S, G SGS. Hybrid LSTM and Ensemble Model for Credit Card Fraud Detection. International Journal of Creative and Open Research in Engineering and Management. 2026;02(05):1-5. doi:10.55041/ijcope.v2i5.537

BibTeX
@article{Saraf2026,
  author = {Saraf, Ruchita Saraf Ruchita and Sanikere, Shravya Sanikere Shravya and G, Tejas N G Tejas N and Shivam, Shivam and G, Savitha G Savitha},
  title = {Hybrid LSTM and Ensemble Model for Credit Card Fraud Detection},
  journal = {International Journal of Creative and Open Research in Engineering and Management},
  year = {2026},
  volume = {02},
  number = {05},
  pages = {1--5},
  doi = {10.55041/ijcope.v2i5.537}
}
Kostenlos starten →