Description Format References Examples
This dataset contains the 40 bootstraps replicate for each of the 17 models. These models have been fitted on the complete dataset!
A list with the 680 outputs of the function CI_all_models
.
This study.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 | ## The names of the models:
names(Models_boot)
## The AIC of model 1 replicate 1:
Models_boot[["mod1_1"]]$AIC
## The convergence status of model 1 replicate 1:
Models_boot[["mod1_1"]]$message
## The estimates from model 1 replicate 1:
data.frame(estimates = Models_boot[["mod1_1"]]$param_matrix[, "value"])
## Studying the effect of potential convergence issues:
table(unlist(lapply(Models_boot, function(mod) mod$convergence))) ## 1 = perfect convergence
models_with_pb <- which(unlist(lapply(Models_boot, function(mod) mod$convergence))!=1)
CI_with_pb <- table_model_averaging(models_list = Models,
models_list_boot = Models_boot)
CI_without_pb <- table_model_averaging(models_list = Models,
models_list_boot = Models_boot[!(names(Models_boot) %in% names(models_with_pb))])
max(abs(100*((CI_with_pb$SE - CI_without_pb$SE) / CI_without_pb$SE)), na.rm = TRUE)
### Conclusion: only 3 fits out of 680 reached maximal time allowed before
### reaching full convergence. Those 3 model fits impact the SE of parameter
### estimates by less than 1.8% in the worst situation, which is negligible.
### We thus chose not to exclude them and base all CI computation on all 680
### fits.
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.