Transforming Time-Varying Volatility Series for Regression

Summary

The document raises a preprocessing question: how to transform a time series with changing volatility before using it as a predictor in a regression model. The author reports trying several scaling methods without obtaining satisfactory stationarity, and asks what approaches are suitable for heteroskedastic data.

No plot details, candidate methods, analysis, or answers are included in the supplied text. It therefore identifies the modeling issue but does not establish a preferred transformation or provide evidence that any method improves stationarity. A useful treatment would first distinguish changes in the series’ level or trend from changes in its conditional variance, since scaling alone may not address both. Any transformation would also need to be evaluated in the context of the regression and its assumptions.

Key ideas

• The question concerns time-varying volatility in a series used as a regression variable.
• The author has tried scaling methods but remains unsatisfied with the results.
• The document offers no specific transformation, analysis, or empirical evidence.
• Level stationarity and variance stationarity are distinct properties to assess.

Tags

# transforming variables


# transforming variables












I am would like to create a regression model with different variables however before using these variables in my regression model I would like to transform the variable in order to make it more stationary .The variable that I would like to transform has time varying volatility . So far I have used different scaling techniques to transform my variable but I have not gotten the most desired results .

This is what my variable looks like plotted

As one can see the variable has time varying volatility , what techniques can I use to better transform my variables specifically variables with time varying volatility

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.