Timeline for Dealing with correlated regressors
Current License: CC BY-SA 2.5
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 9, 2014 at 0:22 | comment | added | kjetil b halvorsen♦ | Standardizing might help interpretability of coefficients. That should be the main reason to standardize! | |
| Oct 14, 2010 at 9:04 | comment | added | Stephan Kolassa | @suncoolsu: yes, standardization helps in polynomial regression - but I would go the whole way and transform the data into appropriate polynomial basis functions. Numerical people have been thinking a lot about stuff like this. | |
| Oct 14, 2010 at 0:14 | comment | added | suncoolsu | I deleted it because I didn't want to confuse people with wrong answer. Probably the moderators brought it up again. | |
| Oct 13, 2010 at 20:06 | history | undeleted | user88 | ||
| Oct 13, 2010 at 18:40 | history | deleted | suncoolsu | ||
| Oct 13, 2010 at 18:39 | history | undeleted | suncoolsu | ||
| Oct 13, 2010 at 18:37 | history | deleted | suncoolsu | ||
| Oct 13, 2010 at 18:28 | comment | added | Brett | Agreed. Centering is useful when entering higher order terms, like polynomial or interaction terms. That doesn't seem to be the case here and will not otherwise help with the problem of correlated predictors. | |
| Oct 13, 2010 at 18:23 | comment | added | suncoolsu | @Brett, a typical example where standardization helps is Polynomial Regression. It is always recommended to standardize the regressors. Standardizing doesn't change the correlation matrix, but makes the var cov matrix (which is now the correl matrix) well behaved (called conditioning by @whuber pointing to the condition number of the matrix, IMHO). | |
| Oct 13, 2010 at 18:15 | comment | added | Brett | Standardizing the variables will not affect the correlations among the independent variables and will not "reduce the effect of correlation" in any way that I can think of with respect to this problem. | |
| Oct 13, 2010 at 18:09 | history | edited | suncoolsu | CC BY-SA 2.5 |
deleted 5 characters in body
|
| Oct 13, 2010 at 18:03 | history | edited | suncoolsu | CC BY-SA 2.5 |
added 17 characters in body
|
| Oct 13, 2010 at 17:57 | comment | added | whuber♦ | Linear transformations (like these) never change correlation coefficients. The point to standardization is to improve the conditioning of the normal matrix. | |
| Oct 13, 2010 at 17:55 | history | edited | suncoolsu | CC BY-SA 2.5 |
added 9 characters in body
|
| Oct 13, 2010 at 17:48 | history | answered | suncoolsu | CC BY-SA 2.5 |