Timeline for Combining regression models based on missing data patterns
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 6, 2024 at 2:51 | comment | added | Squan Schmaan | Thanks, Does Xgboost require the data to be MCAR? In my dataset there are patterns in the missingness (i.e one sensor was not always possible to use depending on a range of factors) | |
| Apr 24, 2024 at 6:57 | comment | added | M. Londschien | I'd advise against 2. Fitting $p$ models, one for each variable using that variable only, and then aggregating their predictions will not allow the models to model interactions between the variables. Even if you use XGBoost. This is not reduced modeling, which also builds an ensemble, but of models for all possible subsets of variables. See my response below. | |
| Apr 24, 2024 at 6:08 | history | bounty awarded | CommunityBot | ||
| Apr 23, 2024 at 23:42 | vote | accept | Squan Schmaan | ||
| Apr 20, 2024 at 14:15 | history | answered | Mark White | CC BY-SA 4.0 |