Small sample fraud detection involves extreme class imbalance and scarce positive instances, thus creating extreme difficulties for typical machine learning paradigms. This work introduces an adaptive regularization boosting framework for boosting algorithms that involves dynamic update rules for weights and theoretical convergence guarantees. The approach introduces a new temperature-calibrated loss function with regularization terms and provides convergence analysis of the proposed framework under small samples. Experimental comparison across five fraud detection data sets shows performance improvements ranging from 5.8% to 15.1% across different datasets with computational tractability preserved. Methodology contributes to ensemble learning by examining boosting behavior in imbalanced settings.

Luqing Ren. Boosting Algorithm Optimization Technology for Ensemble Learning in Small Sample Fraud Detection. Academic Journal of Engineering and Technology Science (2025), Vol. 8, Issue 4: 53-60. https://doi.org/10.25236/AJETS.2025.080407.

[1] Xie, Y., Liu, G., & Yan, C. (2022). An enhanced fraud detection framework for credit card transactions using ensemble machine learning. IEEE Access, 10, 89012-89025.

[2] Taha, A. A., & Malebary, S. J. (2020). An intelligent approach to credit card fraud detection using an optimized light gradient boosting machine. IEEE Access, 8, 25579-25587.

[3] Freund, Y., & Schapire, R. E. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 55(1), 119-139.

[4] Bartlett, P. L., & Mendelson, S. (2002). Rademacher and Gaussian complexities: Risk bounds and structural results. Journal of Machine Learning Research, 3, 463-482.

[5] He, H., & Garcia, E. A. (2009). Learning from imbalanced data. IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263-1284.

[6] Mohri, M., Rostamizadeh, A., & Talwalkar, A. (2018). Foundations of Machine Learning. MIT Press.

[7] Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. Proceedings of the 34th International Conference on Machine Learning, 70, 1321-1330.

[8] Lazarevic, A., & Obradovic, Z. (2002). Boosting algorithms for parallel and distributed learning. Distributed and Parallel Databases, 11(2), 203-229.

[9] Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785-794.

[10] Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189-1232.