Timeline for Why doesn't Random Forest handle missing values in predictors?
Current License: CC BY-SA 3.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 16, 2014 at 14:09 | comment | added | Sycorax♦ | I'm not an expert on GBM, but the RF handling of missing values appears to be rooted in the idea of imputation, en.wikipedia.org/wiki/Imputation_(statistics) In cases where missing values are not missing at random, your results can be biased due to missingness. Imputation attempts to recover this missing values and reduce bias. | |
| May 16, 2014 at 13:22 | comment | added | Fedorenko Kristina | Thank you for your answer! But, both this methods are replacing missing values. But in GBM or regression trees missing values don't replace for anything. What is theoretical difference between, for example GBM and RF in this sense? | |
| May 16, 2014 at 13:19 | history | edited | Sycorax♦ | CC BY-SA 3.0 | added 417 characters in body |
| May 16, 2014 at 13:13 | history | answered | Sycorax♦ | CC BY-SA 3.0 |