Skip to main content
Cross Validated

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

Required fields*

3
  • $\begingroup$ Could you be a little more specific about how you will (a) detect level shifts in such short sequences, (b) exploit the presumption that this is not a level shift but rather a change in variance, (c) exploit the presumption that the variance will initially decrease and then level off, and (d) capitalize on having data from many parallel test procedures. (I haven't seen any examples so far of Autobox or SAS handling any of these special characteristics of the problem, and all of them provide powerful opportunities for better procedures than usually found in time series software.) $\endgroup$ Commented Oct 1, 2012 at 17:18
  • $\begingroup$ @whuber With 10 values as above , one would want to apply the least disruptive remedy. Theree is no doubtthat one could argue that there has been an increase in variance at period 8. The consequences of such a hypothesis going forward could have downsize effects. Untreated pulses I.E. changes in the mean of the errors can be mis-diagnosed as changes in the level, changes in parameters, changes in variance or more simply and less drastic ...just simply unusual i.e. pulses. As more observations become available it will become clearer as to the most probable cause of the exceptional activity. $\endgroup$ Commented Oct 15, 2012 at 16:18
  • $\begingroup$ @whuber The detection of a level shift is accomplishe by Intervention Detection (note this is not the same as Interevntion Modeling) Essentially it is sequence of trial baloons taht are evaluated based upon the effeciveness of the candidate structure. The best way to understand this is to see Tsay unc.edu/~jbhill/tsay.pdf. $\endgroup$ Commented Oct 15, 2012 at 17:37