search

weecology / LDATS

Latent Dirichlet Allocation coupled with Bayesian Time Series analyses
https://weecology.github.io/LDATS
Other
25 stars 5 forks source link

incorporate autocorrelation stucture in change point analysis #19

Open juniperlsimonis opened 6 years ago

juniperlsimonis commented 6 years ago

Lund R, Wang XL, Lu Q, Reeves J, Gallagher C, Feng Y. 2007. Changepoint detection in periodic and autocorrelated time series. Journal of Climate 20: 5178–5190, DOI: 10.1175/JCLI4291.1

juniperlsimonis commented 6 years ago

@davharris and @emchristensen did you all look into incorporating autocorrelation in the change point component? in digging into the literature a bit more, it looks like this is a topic of discussion and work, as it's possible that autocorrelation can impact the determination of changepoints. of particular relevance is that if ac is strong and positive, but not included in the model, it can translate to change points being assigned to times that are just runs of autocorrelated residuals.

davharris commented 6 years ago

Nope. That’s a good idea though.

Dave

On Feb 9, 2018, at 4:39 PM, Juniper Simonis notifications@github.com wrote:

@davharris and @emchristensen did you all look into incorporating autocorrelation in the change point component? in digging into the literature a bit more, it looks like this is a topic of discussion and work, as it's possible that autocorrelation can impact the determination of changepoints. of particular relevance is that if ac is strong and positive, but not included in the model, it can translate to change points being assigned to times that are just runs of autocorrelated residuals.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

juniperlsimonis commented 6 years ago

roger, thanks for letting me know!