Change-Point Detection in Time-Series Data by Relative Density-Ratio Estimation

Abstract

The objective of change-point detection is to discover abrupt property changes lying behind time-series data. In this paper, we present a novel statistical change-point detection algorithm based on non-parametric divergence estimation between time-series samples from two retrospective segments. Our method uses the relative Pearson divergence as a divergence measure, and it is accurately and efficiently estimated by a method of direct density-ratio estimation. Through experiments on artificial and real-world datasets including human-activity sensing, speech, and Twitter messages, we demonstrate the usefulness of the proposed method.

References in corpus (3)

• Estimating divergence functionals and the likelihood ratio by convex risk minimization
• Bayesian Online Changepoint Detection
• Condition Number Analysis of Kernel-based Density Ratio Estimation

Cited by in corpus (39)

• Selective review of offline change point detection methods
• A Survey of Deep Learning-based Object Detection
• Time Series Change Point Detection with Self-Supervised Contrastive Predictive Coding
• Detecting Regions of Maximal Divergence for Spatio-Temporal Anomaly Detection
• Adaptive, locally-linear models of complex dynamics
• Change Point Detection in Time Series Data using Autoencoders with a Time-Invariant Representation
• Entropic Dynamic Time Warping Kernels for Co-evolving Financial Time Series Analysis
• ESPRESSO: Entropy and ShaPe awaRe timE-Series SegmentatiOn for processing heterogeneous sensor data
• TweetsKB: A Public and Large-Scale RDF Corpus of Annotated Tweets
• Time Series Segmentation through Automatic Feature Learning
• A Note on Kernel Methods for Multiscale Systems with Critical Transitions
• Optimal nonparametric change point detection and localization
• Characterizing Driving Context from Driver Behavior
• Continual Learning in Practice
• Online Neural Networks for Change-Point Detection
• Deep density ratio estimation for change point detection
• Deep Learning for Multi-Scale Changepoint Detection in Multivariate Time Series
• Direct Density-Derivative Estimation and Its Application in KL-Divergence Approximation
• Detecting Change Intervals with Isolation Distributional Kernel
• Stein Variational Online Changepoint Detection with Applications to Hawkes Processes and Neural Networks
• Latent Neural Stochastic Differential Equations for Change Point Detection
• Meta-Learning for Relative Density-Ratio Estimation
• Trend and Variance Adaptive Bayesian Changepoint Analysis & Local Outlier Scoring
• Generalization of Change-Point Detection in Time Series Data Based on Direct Density Ratio Estimation
• Model Monitoring and Dynamic Model Selection in Travel Time-series Forecasting
• Semantic Analysis of Traffic Camera Data: Topic Signal Extraction and Anomalous Event Detection
• Understanding Action Sequences based on Video Captioning for Learning-from-Observation
• Non-parametric Structural Change Detection in Multivariate Systems
• Structure Learning of Partitioned Markov Networks
• Maximally Divergent Intervals for Anomaly Detection
• ReAD: A Regional Anomaly Detection Framework Based on Dynamic Partition
• Post Selection Inference with Incomplete Maximum Mean Discrepancy Estimator
• Hierarchical Quickest Change Detection via Surrogates
• Block-Wise MAP Inference for Determinantal Point Processes with Application to Change-Point Detection
• SAFE: Spectral Evolution Analysis Feature Extraction for Non-Stationary Time Series Prediction
• Shape-CD: Change-Point Detection in Time-Series Data with Shapes and Neurons
• Process Knowledge Driven Change Point Detection for Automated Calibration of Discrete Event Simulation Models Using Machine Learning
• Shedding Light on the Targeted Victim Profiles of Malicious Downloaders
• Continual Density Ratio Estimation in an Online Setting