Detection of trend changes in time series using Bayesian inference

TL;DR

This paper presents a Bayesian inference method for detecting change points in time series data, providing confidence intervals and validated on synthetic and real-world Nile River flow data.

Contribution

The work introduces a Bayesian approach for change point detection that estimates their locations and confidence intervals, improving sensitivity and reliability over existing methods.

Findings

Successfully detects change points in synthetic data.

Identifies a significant transition in Nile River flow data.

Provides confidence intervals for estimated change points.

Abstract

Change points in time series are perceived as isolated singularities where two regular trends of a given signal do not match. The detection of such transitions is of fundamental interest for the understanding of the system's internal dynamics. In practice observational noise makes it difficult to detect such change points in time series. In this work we elaborate a Bayesian method to estimate the location of the singularities and to produce some confidence intervals. We validate the ability and sensitivity of our inference method by estimating change points of synthetic data sets. As an application we use our algorithm to analyze the annual flow volume of the Nile River at Aswan from 1871 to 1970, where we confirm a well-established significant transition point within the time series.