Timeline for Log or square-root transformation for ARIMA
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 9 at 13:19 | comment | added | whuber♦ | @love This issue is discussed extensively here on CV. Search our site for answers about regression intercept, for instance. | |
| Mar 20, 2023 at 7:36 | comment | added | lovetl2002 |
In the fit <- lm(log(z.spread) ~ log(z.med)), should I include the constant or not?
|
|
| Mar 14, 2023 at 15:53 | comment | added | whuber♦ | @lovetl2002 Certainly! That is one of the original uses of Lowess. | |
| Mar 14, 2023 at 11:00 | comment | added | lovetl2002 | Can I use Lowess to get the level and subtract it from y (then use abs) to get the spread? | |
| Jul 22, 2020 at 14:48 | comment | added | whuber♦ | @Stats It's ok, but one of the virtues of the Box-Cox transformation (instead of the straight power transformation) is that it does not reverse trends. The Box-Cox analog of your transformation is $x\to (x^{-0.87}-1)/(-0.87).$ Notice how the division by the (negative) power reverses the trend back again. See stats.stackexchange.com/a/467525/919 for more discussion. | |
| Jul 22, 2020 at 14:10 | comment | added | StatsMonkey | @whuber I did power transformation on a ts and got -0.870478 as value of lambda (since slope > 1). After applying it, the trend is reversed - Is this (negative power and trend reversal) okay while stabilizing the variance? | |
| Nov 5, 2013 at 18:27 | comment | added | whuber♦ | @Nick I agree that the trends (pre-2004 and post-2004 separately) are more linear for the logs. Some method to handle the heteroscedasticity would be essential for making forecasts: without that, forecasts based on the untransformed data would have optimistically narrow prediction intervals whereas forecasts based on logs would have pessimistically wide intervals. A short-term forecast may be robust to deviations from linearity--making cube roots a good first choice--but making longer-term forecasts does require close attention to the trend, initially favoring the log as you claim. | |
| Nov 5, 2013 at 13:09 | comment | added | digdeep | @whuber thanks for the detailed write up. Really helpful post and some neat R code too. | |
| Nov 4, 2013 at 23:47 | comment | added | Nick Cox | This is very nicely done as always. I would put a bit more emphasis on getting trend right first and a bit less on getting the handling of variability. That boosts the case for logarithms. No bias against cube roots, which feature in various papers of mine. | |
| Nov 4, 2013 at 22:59 | history | answered | whuber♦ | CC BY-SA 3.0 |