Message Boards Message Boards

Finding all structural breaks in time series

Anton Antonov, Accendo Data LLC
Posted 5 years ago
POSTED BY: Anton Antonov
25 Replies
Sort By:
Erfan Abdi
Erfan Abdi
Posted 1 year ago

Thanks for your time and waiting for your reply
POSTED BY: Erfan Abdi

I will have time to take a look after two days. (Sorry for the delay...)
POSTED BY: Anton Antonov
Erfan Abdi
Erfan Abdi
Posted 1 year ago

Thank you very much. My time axis is months (288 months from 2000 to 2023). How can I modify the following command to show (170, 200, 240, 255) in the Chow diagram as shown below?

enter image description here

enter image description here
POSTED BY: Erfan Abdi

Thank you for your interest in this paclet!

What is the interpretation of the time axis -- weeks, days? The time-values are uniform -- are they "just" observation indexes?

I think too many knots are used. (In the attached notebook you provided.) I think the knots should be [20,25].

Also, if we look into the regression quantiles fits:

qrObj2 =
  QRMonUnit[data]⟹
   QRMonEchoDataSummary⟹
   QRMonQuantileRegression[20, {0.1, 0.5, 0.9}, InterpolationOrder -> 2]⟹
   QRMonPlot[ImageSize -> Large, AspectRatio -> 1/3, GridLines -> {{170, 200, 240, 255}, None}];

maybe the structural breaks are just at {170, 200, 240, 255}.

enter image description here
POSTED BY: Anton Antonov
Erfan Abdi
Erfan Abdi
Posted 1 year ago

Hello, Dr. Antonov, have a good day. I have done your package for monthly wind speed temporal data (period 2000-2023), please let me know your opinion on its accuracy

Attachments:
wind speed-(Dr.A...nb
POSTED BY: Erfan Abdi

I forgot to put the paclet installation command in the notebook -- I just updated it. Please, use that one.
POSTED BY: Anton Antonov
Alex Teymouri
Alex Teymouri
Posted 1 year ago

Thank you so much, Anton.

I never forget your help.
POSTED BY: Alex Teymouri

Thanks for trying out those functionalities!

Please make sure:

  • All data elements are numeric
  • Correct syntax pipeline symbol is used between the monadic operations : "⇒" instead of "?"

Remark: Apparently there is a "bug" in the post itself -- it shows code with "?" instead of "⇒".

See the attached notebook.

Attachments:
Structural break...nb
POSTED BY: Anton Antonov
Alex Teymouri
Alex Teymouri
Posted 1 year ago

Hi Anton,

Thank you for sharing the excellent package. How do I determine the year of the breaking point? I am a beginner in Mathematica. sorry.

data = {{"1989", 191.27}, {"1990", 148.01`}, {"1991", 
   249.13}, {"1992", 277.84}, {"1993", 362.4`}, {"1994", 
   374.18}, {"1995", 175.56`}, {"1996", 251.49}, {"1997", 
   192.92}, {"1998", 233.06`}, {"1999", 220.79}, {"2000", 
   204.91}, {"2001", 203.68`}, {"2002", 302.94`}, {"2003", 
   218.92`}, {"2004", 284.76`}, {"2005", 233.30}, {"2006", 
   305.13`}, {"2007", 229.95`}, {"2008", 171.53}, {"2009", 
   241.91`}, {"2010", 183.92}, {"2011", 282.23`}, {"2012", 
   217.19}, {"2013", 262.53}, {"2014", 310.97`}, {"2015", 
   293.45`}, {"2016", 290.96}, {"2017", 204.36`}, {"2018", 
   388.73}, {"2019", 264.99`}, {"2020", 345.66`}}

enter image description here
POSTED BY: Alex Teymouri
POSTED BY: Anton Antonov
Murtaza Dar
Murtaza Dar
Posted 5 years ago

how can I get structure break by time-series data of river runoff? which software should I use to get a structured break?
POSTED BY: Murtaza Dar

Please see the last 20-22 minutes of the second video of the Quantile Regression live-coding sessions.
POSTED BY: Anton Antonov

Hi, Thanks to Dr. Antonov, I can't find the youtube video of the definition of this structural break?
POSTED BY: Nasrin Attar
POSTED BY: Anton Antonov

Where is it?

You forgot to attach the older version.
POSTED BY: M.A. Ghorbani

I changed the implementation to work with older Mathematica versions -- please try it out.

POSTED BY: Anton Antonov

You right. I have a version 11.1.

Indeed I wanted to invite you to this study as a collaborative study and apply your package for the 222 stations record. It was an honor to me if I had this opportunity.

Thank you so much for everything. Mohammad
POSTED BY: M.A. Ghorbani

It looks like you are using an older version of Mathematica, not the current one, Version 12.0. What is your Mathematica version?

Fit in Mathematica 12.0 can take 4 arguments: see the attached screenshot.

enter image description here
POSTED BY: Anton Antonov
Attachments:
Dr Antonov.nb
POSTED BY: M.A. Ghorbani

Thanks for trying out that package!

The error message you got:

MonadicQuantileRegressionPrivateChowTestStatistic::empdata: The split point 1.` produced an empty split dataset.

indicates what is the problem: you are using the min x-value as splitting point. If you use splitting points away from min and max x-values the computations proceed normally; see the attached screenshot.

enter image description here
POSTED BY: Anton Antonov

Dear Antonov,

I am very interested in learning structural breaks in the time series. I started to apply your package for the rainfall time series. Please see the enclosed notebook. I got an error. What can I do now?

I have 222 rain gauge stations over Turkey, so I should solve the error so I can analysis for other stations.

Thanks for your time.

Regards, Mohammad

Attachments:
Rainfall.nb
POSTED BY: M.A. Ghorbani

Thank you, Staff Team!
POSTED BY: Anton Antonov

enter image description here - Congratulations! This post is now featured in our Staff Pick column as distinguished by a badge on your profile of a Featured Contributor! Thank you, keep it coming, and consider contributing your work to the The Notebook Archive!

POSTED BY: EDITORIAL BOARD

Thank you for your note!

You might want to follow this dedicated MathematicaVsR project, in which I plan to discuss various algorithms and measures for finding anomalies in time series. (I plan to give a presentation at WTC-2019 on that subject.)
POSTED BY: Anton Antonov
Jason Zhao
Jason Zhao, GEICO
Posted 5 years ago

This is excellent! Your post solves my question in a long ago post.
POSTED BY: Jason Zhao
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Tag limit exceeded
Note: Only the first five people you tag will receive an email notification; the other tagged names will appear as links to their profiles.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract