Multiple Changepoint Detection via Genetic Algorithms

This paper studies genetic algorithms as a means of estimating the number of changepoints and their locations in a climatic time series. Such methods bypass classic subsegmentation algorithms, which sometimes yield suboptimal conclusions. Minimum description length techniques are introduced. These t...

Full description

Saved in:
Bibliographic Details
Published in:Journal of climate 2012-01, Vol.25 (2), p.674-686
Main Authors: Li, Shanghong, Lund, Robert
Format: Article
Language:English
Subjects:
Algorithms
Annual precipitation
Climate change
Cyclones
Data collection
Earth, ocean, space
Exact sciences and technology
External geophysics
Fines & penalties
Genetic algorithms
Hurricanes
Information theory
Integers
Marine
Mathematical minima
Mathematical models
Methods
Objective function
Objective functions
Optimization
Parametric models
Precipitation
Precipitation measurements
Series convergence
Studies
Time series
Trends
Tropical cyclones
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper studies genetic algorithms as a means of estimating the number of changepoints and their locations in a climatic time series. Such methods bypass classic subsegmentation algorithms, which sometimes yield suboptimal conclusions. Minimum description length techniques are introduced. These techniques require optimizing an objective function over all possible changepoint numbers and location times. The general objective functions allow for correlated data, reference station aspects, and/or nonnormal marginal distributions, all common features of climate time series. As an exhaustive evaluation of all changepoint configurations is not possible, the optimization is accomplished via a genetic algorithm that randomly walks through a subset of good models in an intelligent manner. The methods are applied in the analysis of 173 yr of annual precipitation measurements from New Bedford, Massachusetts, and the North Atlantic basin tropical cyclone record.
ISSN:0894-8755
1520-0442
DOI:10.1175/2011JCLI4055.1