Permutation tests in change point analysis

Antoch, Jaromı́r; Hušková, Marie

doi:10.1016/s0167-7152(01)00009-8

Cited by 64 publications

(43 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Let H 0 be the null hypothesis that there is no change in the jellyfish spatial distribution over the study period, whereas the alternative states that the spatial distribution changes to a different distribution after a threshold year. See, also, Antoch and Huskova (2001) and Hu, Qin, and Follmann (2008). Under H 0 , each set of data with the year values randomly shuffled is equally likely to occur, given everything else being equal, so that the observed maximum penalized log-likelihood is distributed according to the distribution of the maximum penalized log-likelihood of the randomly shuffled data.…”

Section: Methodsmentioning

confidence: 99%

Nonparametric Threshold Model of Zero-Inflated Spatio-Temporal Data with Application to Shifts in Jellyfish Distribution

Liu

Ciannelli

Decker

et al. 2010

JABES

View full text Add to dashboard Cite

There is increasing scientific interest in studying the spatial distribution of species abundance in relation to environmental variability. Jellyfish in particular have received considerable attention in the literature and media due to regional population increases and abrupt changes in distribution. Jellyfish distribution and abundance data, like many biological datasets, are characterized by an excess of zero counts or nonstationary processes, which hampers their analyses by standard statistical methods. Here we further develop a recently proposed statistical framework, the constrained zero-inflated generalized additive model (COZIGAM), and apply it to a spatio-temporal dataset of jellyfish biomass in the Bering Sea. Our analyses indicate systematic spatial variation in the process that causes the zero inflation. Moreover, we show strong evidence of a range expansion of jellyfish from the southeastern to the northwestern portion of the survey area beginning in 1991. The proposed methodologies could be readily applied to ecological data in which zero inflation and spatio-temporal nonstationarity are suspected, such as data describing species distribution in relation to changes of climate-driven environmental variables. Some supplemental materials including an animation of jellyfish annual biomass and web appendices are available online.

show abstract

Section: Methodsmentioning

confidence: 99%

Nonparametric Threshold Model of Zero-Inflated Spatio-Temporal Data with Application to Shifts in Jellyfish Distribution

Liu

Ciannelli

Decker

et al. 2010

JABES

View full text Add to dashboard Cite

show abstract

“…More precisely, Antoch and Hušková (2001) show that the untrimmed test statistic sup π∈(0,1) |Z π | -i.e., with Π = (0, 1) -has an extreme value distribution that is denoted D 0 ∞|Y in the following. Analogously to the ideas above, the corresponding exact conditional distribution D 0 σ|Y can again be approximated by P draws from the permutation distribution of the untrimmed statistic.…”

Section: Distribution Of the Test Statisticmentioning

confidence: 99%

“…However, several applications exist, see e.g., Kennedy (1995) for an overview on how to employ permutation tests in econometrics. For the specific problem of structural change, permutation tests are not typically considered in econometrics, whereas in the statistics literature they are used more frequently, especially in nonparametric tests based on maximally selected rank statistics (Lausen and Schumacher 1992;Hothorn and Lausen 2003;Boulesteix and Strobl 2007) and in mathematical statistics (Antoch and Hušková 2001;Kirch and Steinebach 2006;Kirch 2007). Building on these ideas, we discuss in the following how a wide class of permutation tests for structural change can be established, pointing out their strengths and weaknesses.…”

Section: Introductionmentioning

confidence: 99%

A toolbox of permutation tests for structural change

Zeileis

Hothorn

2013

Stat Papers

View full text Add to dashboard Cite

The supLM test for structural change is embedded into a permutation test framework for a simple location model. The resulting conditional permutation distribution is compared to the usual (unconditional) asymptotic distribution, showing that the power of the test can be clearly improved in small samples. Furthermore, the permutation test is embedded into a general framework that encompasses tools for binary and multivariate dependent variables as well as model-based permutation testing for structural change. It is also demonstrated that the methods can not only be employed for analyzing structural changes in time series data but also for recursive partitioning of cross section data. The procedures suggested are illustrated using both artificial data and empirical applications (number of youth homicides, employment discrimination data, carbon flux in tropical forests, stock returns, and demand for economics journals).

show abstract

“…For a thorough introduction we refer to Good (2005). In change-point analysis permutation methods were first suggested by Antoch and Hušková (2001) and later pursued by others (for a recent survey confer Hušková 2004). Berkes et al (2004) showed that the bootstrap provides better approximations for the critical values than asymptotics in a number of change-point situations.…”

mentioning

confidence: 99%

Bootstrapping sequential change-point tests for linear regression

Hušková

Kirch

2011

Metrika

Self Cite

View full text Add to dashboard Cite

Bootstrap methods for sequential change-point detection procedures in linear regression models are proposed. The corresponding monitoring procedures are designed to control the overall significance level. The bootstrap critical values are updated constantly by including new observations obtained from the monitoring. The theoretical properties of these sequential bootstrap procedures are investigated, showing their asymptotic validity. Bootstrap and asymptotic methods are compared in a simulation study, showing that the studentized bootstrap tests hold the overall level better especially for small historic sample sizes while having a comparable power and run length.

show abstract

Permutation tests in change point analysis

Cited by 64 publications

References 3 publications

Nonparametric Threshold Model of Zero-Inflated Spatio-Temporal Data with Application to Shifts in Jellyfish Distribution

Nonparametric Threshold Model of Zero-Inflated Spatio-Temporal Data with Application to Shifts in Jellyfish Distribution

A toolbox of permutation tests for structural change

Bootstrapping sequential change-point tests for linear regression

Contact Info

Product

Resources

About