An Interpretable Generative Framework for Anomaly Detection in High-Dimensional Financial Time Series

TL;DR

This paper introduces ReGEN-TAD, an interpretable generative model combining machine learning and econometrics to detect anomalies in high-dimensional financial time series, improving robustness and interpretability.

Contribution

It presents a novel framework that integrates forecasting, reconstruction, and econometric diagnostics for anomaly detection without labeled data.

Findings

Enhanced robustness to structured deviations

Economically coherent factor-level attribution

Unified anomaly scoring method

Abstract

Detecting structural instability and anomalies in high-dimensional financial time series is challenging due to complex temporal dependence and evolving cross-sectional structure. We propose ReGEN-TAD, an interpretable generative framework that integrates modern machine learning with econometric diagnostics for anomaly detection. The model combines joint forecasting and reconstruction within a refined convolutional--transformer architecture and aggregates complementary signals capturing predictive inconsistency, reconstruction degradation, latent distortion, and volatility shifts. Robust calibration yields a unified anomaly score without labeled data. Experiments on synthetic and financial panels demonstrate improved robustness to structured deviations while enabling economically coherent factor-level attribution.