Moving sum procedure for multiple change point detection in large factor models

2024-10-03

AI Summary

This paper addresses the detection of multiple structural breaks in high-dimensional time series factor models, focusing on two mechanisms: abrupt changes in factor loadings and the emergence/disappearance of latent factors. We propose the first moving sum (MOSUM) test specifically designed for large-dimensional factor models—extending MOSUM to high-dimensional settings with latent factors for the first time. We establish asymptotic theory under family-wise error rate (FWER) control: under the null hypothesis, the test statistic converges to an extreme-value distribution; under alternatives, break-point estimators are consistent. Multiple testing correction ensures reliable inference. Monte Carlo simulations and empirical analysis on large-scale financial volatility data demonstrate that our method significantly outperforms existing approaches in detection accuracy, localization precision, and simultaneous control of false positives.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Multiagent Systems: Mechanism Design

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience

Abstract

The paper proposes a moving sum methodology for detecting multiple change points in high-dimensional time series under a factor model, where changes are attributed to those in loadings as well as emergence or disappearance of factors. We establish the asymptotic null distribution of the proposed test for family-wise error control, and show the consistency of the procedure for multiple change point estimation. Simulation studies and an application to a large dataset of volatilities demonstrate the competitive performance of the proposed method.

Problem

• Detect multiple change points in high-dimensional time series
• Control family-wise error in factor model change detection
• Estimate changes in loadings and factor emergence/disappearance

Innovation

• Moving sum method for change point detection
• Handles high-dimensional factor model changes
• Asymptotic null distribution for error control