Anomaly detection by partitioning of multi-variate time series

2025-09-22

AI Summary

In multivariate time series anomaly detection, global modeling often overlooks local correlation structures, leading to insufficient sensitivity and robustness. To address this, we propose an unsupervised, correlation-driven variable partitioning framework. First, we compute the Pearson correlation matrix among variables and apply spectral clustering to automatically identify cohesive, weakly coupled variable subsets. Then, classical anomaly detectors—e.g., AutoEncoder or Isolation Forest—are independently deployed on each subset, enabling localized modeling while preserving intrinsic variable dependencies. The method is fully unsupervised, distribution-agnostic, and inherently parallelizable. Extensive experiments on multiple synthetic and real-world benchmark datasets demonstrate that our approach achieves an average 8.2% improvement in F1-score over state-of-the-art methods, with superior robustness to noise and sparse anomalies.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Time-Series/Data Streams
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

In this article, we suggest a novel non-supervised partition based anomaly detection method for anomaly detection in multivariate time series called PARADISE. This methodology creates a partition of the variables of the time series while ensuring that the inter-variable relations remain untouched. This partitioning relies on the clustering of multiple correlation coefficients between variables to identify subsets of variables before executing anomaly detection algorithms locally for each of those subsets. Through multiple experimentations done on both synthetic and real datasets coming from the literature, we show the relevance of our approach with a significant improvement in anomaly detection performance.

Problem

• Detecting anomalies in multivariate time series data
• Partitioning variables while preserving inter-variable relationships
• Improving anomaly detection performance through local analysis

Innovation

• Partitioning variables while preserving inter-variable relations
• Clustering correlation coefficients to identify variable subsets
• Executing anomaly detection locally on each subset