RGLD: Randomized Global-Local Density Estimation for Tabular Anomaly Detection

2026-06-27

AI Summary

This work addresses the challenge of simultaneously achieving high accuracy, cross-dataset robustness, and computational efficiency in unsupervised tabular anomaly detection by proposing a randomized global–local density estimation method. The approach uniquely integrates global low-density region identification with local neighborhood support analysis, jointly modeling anomalies across multiple random views generated via feature bagging to uncover latent anomalous signals obscured in any single representation. Extensive experiments on 47 benchmark datasets demonstrate that the method significantly outperforms 23 baseline approaches, achieving the best AUROC and the second-best AUPRC scores. Moreover, its inference speed is 50 to 580 times faster than deep learning–based methods while approaching the computational efficiency of traditional statistical techniques.

Abstract

Unsupervised tabular anomaly detection requires methods that are accurate, robust across heterogeneous datasets, and computationally efficient. Classical statistical detectors are often efficient, but they usually rely on a fixed data view and a single notion of abnormality. Deep anomaly detectors can learn more flexible scoring functions, but they are substantially slower and difficult to tune in unsupervised settings due to the lack of a reliable supervisory signal. We propose RGLD, a randomized global-local density estimator for efficient unsupervised tabular anomaly detection. RGLD combines a global random-feature density branch, which identifies samples in broadly low-density regions, with a local neighbor branch, which detects samples that are weakly supported by nearby observations. Both branches operate over feature-bagged randomized views, allowing RGLD to expose anomaly evidence that may be hidden in any single representation. We conduct experiments on 47 tabular datasets against 23 statistical and deep anomaly detection baselines under fully unsupervised setting. RGLD achieves the strongest dataset-level AUROC performance, ranking 1st in dataset wins, and ranks 2nd in AUPRC wins. RGLD is also faster than all evaluated deep detectors, achieving 50x-580x speedups, and remains competitive with statistical methods in runtime, yielding a favorable accuracy-efficiency tradeoff.

Problem

• unsupervised anomaly detection
• tabular data
• density estimation
• computational efficiency
• heterogeneous datasets

Innovation

• Randomized Density Estimation
• Global-Local Anomaly Detection
• Unsupervised Tabular Anomaly Detection
• Feature Bagging
• Efficient Deep Anomaly Detection