Financial Data Analysis with Robust Federated Logistic Regression

TL;DR

The paper tackles privacy-preserving financial data analysis by introducing a robust federated logistic regression (FLR) framework. It leverages three aggregation strategies—mean, median, and trimmed mean—to mitigate outlier effects and maintain interpretability intrinsic to logistic regression. Through experiments on four public financial datasets, FLR demonstrates competitive AUC performance against centralized baselines under both IID and non-IID data, with robustness to outliers, while preserving model transparency via interpretable coefficients. The work emphasizes practical impact for financial decision-making with distributed data, and provides evidence that robust FLR can match traditional methods without sharing raw data. Overall, the approach offers a scalable, interpretable, and privacy-conscious alternative for financial risk analysis in federated settings.

Abstract

In this study, we focus on the analysis of financial data in a federated setting, wherein data is distributed across multiple clients or locations, and the raw data never leaves the local devices. Our primary focus is not only on the development of efficient learning frameworks (for protecting user data privacy) in the field of federated learning but also on the importance of designing models that are easier to interpret. In addition, we care about the robustness of the framework to outliers. To achieve these goals, we propose a robust federated logistic regression-based framework that strives to strike a balance between these goals. To verify the feasibility of our proposed framework, we carefully evaluate its performance not only on independently identically distributed (IID) data but also on non-IID data, especially in scenarios involving outliers. Extensive numerical results collected from multiple public datasets demonstrate that our proposed method can achieve comparable performance to those of classical centralized algorithms, such as Logistical Regression, Decision Tree, and K-Nearest Neighbors, in both binary and multi-class classification tasks.

Figures (20)

• Figure 1: Feature correlation of CreditRisk after iteratively removing the features with higher VIFs (>10).
• Figure 2: The effect of different sampling sizes $s$ with $M=100$ clients and $p_{out}=0.1$ on non-IID data.
• Figure 3: The effect of number of clients $M$ with fixed $p_{out}=0.1$ and $s=100$ on non-IID data.
• Figure 4: The effect of number of clients $M$ with fixed $p_{out}=0.1$ and $s=100$ on non-IID data.
• Figure 6: Poor vs. Rest