Timeline for High $R^2$ on Ordinary least squares model with violated assumptions. Is it good?
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 22, 2015 at 9:31 | comment | added | Richard Hardy | OK, that may very well be the case of fitting one random walk against another which is not permitted under OLS assumptions. You may refer to spurious regression and perhaps cointegration. | |
| Dec 22, 2015 at 9:14 | comment | added | marcodena | @whuber I tried removing observation 249 and 381 (which is kinda outlier when I remove 249) and I have r-squared 0.972 | |
| Dec 22, 2015 at 9:09 | comment | added | marcodena | @RichardHardy it's about GDP and population of some cities. | |
| Dec 22, 2015 at 9:08 | comment | added | marcodena | @Repmat Causal effect (?) and fit... | |
| Dec 21, 2015 at 13:11 | comment | added | whuber♦ | What happens to $R^2$ when you remove observation 249 from the dataset and refit the models? | |
| Dec 21, 2015 at 10:10 | history | edited | Richard Hardy | CC BY-SA 3.0 |
edited body
|
| Dec 21, 2015 at 10:09 | comment | added | Richard Hardy | What kind of data is that? By any chance, are you regressing one random walk on another? | |
| Dec 21, 2015 at 9:28 | history | edited | Tim | CC BY-SA 3.0 |
added 4 characters in body; edited title
|
| Dec 21, 2015 at 9:27 | comment | added | Tim | See stats.stackexchange.com/questions/13314/… | |
| Dec 21, 2015 at 9:13 | comment | added | Repmat | What do you mean by best? Causal effects? Prediction? Fit? The $R^2$ is very poor indication of any of these "bests"... | |
| Dec 21, 2015 at 9:05 | history | asked | marcodena | CC BY-SA 3.0 |