Bayesian Online Changepoint Detection

Abstract

Changepoints are abrupt variations in the generative parameters of a data sequence. Online detection of changepoints is useful in modelling and prediction of time series in application areas such as finance, biometrics, and robotics. While frequentist methods have yielded online filtering and prediction techniques, most Bayesian papers have focused on the retrospective segmentation problem. Here we examine the case where the model parameters before and after the changepoint are independent and we derive an online algorithm for exact inference of the most recent changepoint. We compute the probability distribution of the length of the current ``run,'' or time since the last changepoint, using a simple message-passing algorithm. Our implementation is highly modular so that the algorithm may be applied to a variety of types of data. We illustrate this modularity by demonstrating the algorithm on three different real-world data sets.

7 pages, 4 figures, latex

Cited by in corpus (82)

• Selective review of offline change point detection methods
• Change-Point Detection in Time-Series Data by Relative Density-Ratio Estimation
• Change Point Detection in Time Series Data using Autoencoders with a Time-Invariant Representation
• Real-Time Anomaly Detection for Streaming Analytics
• Unsupervised Anomaly Detection in Stream Data with Online Evolving Spiking Neural Networks
• Generalized Variational Inference: Three arguments for deriving new Posteriors
• A taxonomy of surprise definitions
• A Survey on IoT Ground Sensing Systems for Early Wildfire Detection: Technologies, Challenges and Opportunities
• An empirical evaluation of active inference in multi-armed bandits
• Unravelling intermittent features in single particle trajectories by a local convex hull method
• Thompson Sampling in Switching Environments with Bayesian Online Change Point Detection
• Time-Varying Gaussian Process Bandit Optimization
• Edge AI: On-Demand Accelerating Deep Neural Network Inference via Edge Computing
• Neuroscience-Inspired Algorithms for the Predictive Maintenance of Manufacturing Systems
• Time Series Segmentation through Automatic Feature Learning
• Deep Reinforcement Learning amidst Lifelong Non-Stationarity
• Slow Momentum with Fast Reversion: A Trading Strategy Using Deep Learning and Changepoint Detection
• Online Bayesian Passive-Aggressive Learning
• Sparse Multi-Output Gaussian Processes for Medical Time Series Prediction
• WATCH: Wasserstein Change Point Detection for High-Dimensional Time Series Data
• Continual Learning in Practice
• Spatio-temporal Bayesian On-line Changepoint Detection with Model Selection
• Semi-metric portfolio optimization: a new algorithm reducing simultaneous asset shocks
• Bayesian Online Prediction of Change Points
• Explicit-Duration Markov Switching Models
• Scalable Gaussian Processes for Characterizing Multidimensional Change Surfaces
• Causal Discovery from Heterogeneous/Nonstationary Data with Independent Changes
• Semantic Edge Computing and Semantic Communications in 6G Networks: A Unifying Survey and Research Challenges
• Sequential Gaussian Processes for Online Learning of Nonstationary Functions
• Deep Learning for Multi-Scale Changepoint Detection in Multivariate Time Series
• Statistically Significant Detection of Linguistic Change
• Equivalence relations and $L^p$ distances between time series with application to the Black Summer Australian bushfires
• Continuous Meta-Learning without Tasks
• Powerpropagation: A sparsity inducing weight reparameterisation
• Passive detection of behavioral shifts for suicide attempt prevention
• Discovery and Visualization of Nonstationary Causal Models
• Non-Negative Bregman Divergence Minimization for Deep Direct Density Ratio Estimation
• Stein Variational Online Changepoint Detection with Applications to Hawkes Processes and Neural Networks
• LinXGBoost: Extension of XGBoost to Generalized Local Linear Models
• Latent Neural Stochastic Differential Equations for Change Point Detection
• Autoencoder Based Iterative Modeling and Multivariate Time-Series Subsequence Clustering Algorithm
• Bayesian Online Change Point Detection for Baseline Shifts
• Trend and Variance Adaptive Bayesian Changepoint Analysis & Local Outlier Scoring
• Deep Explicit Duration Switching Models for Time Series
• Leveraging Multimodal Haptic Sensory Data for Robust Cutting
• Non-smooth Bayesian Optimization in Tuning Problems
• Efficacy of Statistical and Artificial Intelligence-based False Information Cyberattack Detection Models for Connected Vehicles
• Detecting and Adapting to Irregular Distribution Shifts in Bayesian Online Learning
• Changepoint Detection in the Presence of Outliers
• Bootstrap for change point detection
• Fast likelihood-based change point detection
• Self-Supervised Transformer Architecture for Change Detection in Radio Access Networks
• Bandit Quickest Changepoint Detection
• Active multi-fidelity Bayesian online changepoint detection
• Online Robust and Adaptive Learning from Data Streams
• Towards automated symptoms assessment in mental health
• Anomaly detection and motif discovery in symbolic representations of time series
• Balancing New Against Old Information: The Role of Surprise in Learning
• Bayesian Inference for High Dimensional Changing Linear Regression with Application to Minnesota House Price Index Data
• When and Whom to Collaborate with in a Changing Environment: A Collaborative Dynamic Bandit Solution
• Empirical Bayesian analysis of simultaneous changepoints in multiple data sequences
• Robust seasonal cycle of Arctic sea ice area through tipping point in amplitude
• Hierarchical Quickest Change Detection via Surrogates
• Bayesian Changepoint Analysis
• Dynamic time series clustering via volatility change-points
• Online Discrimination of Nonlinear Dynamics with Switching Differential Equations
• Block-Wise MAP Inference for Determinantal Point Processes with Application to Change-Point Detection
• Quantifying and Reducing Bias in Maximum Likelihood Estimation of Structured Anomalies
• Correlation-aware Unsupervised Change-point Detection via Graph Neural Networks
• A Review of Changepoint Detection Models
• Probabilistic sequential matrix factorization
• On Posterior consistency of Bayesian Changepoint models
• Detecting Change Signs with Differential MDL Change Statistics for COVID-19 Pandemic Analysis
• Confirmatory Bayesian Online Change Point Detection in the Covariance Structure of Gaussian Processes
• Changepoint Analysis of Topic Proportions in Temporal Text Data
• Context selectivity with dynamic availability enables lifelong continual learning
• On The Calibration of Short-Term Interest Rates Through a CIR Model
• Graceful Forgetting II. Data as a Process
• Quickest change detection with unknown parameters: Constant complexity and near optimality
• Simultaneous quantification and changepoint detection of point source gas emissions using recursive Bayesian inference
• Sequential online prediction in the presence of outliers and change points: an instant temporal structure learning approach
• Hidden Markov Models Applied To Intraday Momentum Trading With Side Information