Custom Weighting in ensemble_weighted Function with Multiple Models

[dmresearch15]

Hi,

I am referencing the example from this article on autoregressive machine learning using m4_monthly data.

Instead of using fixed loadings like loadings = c(4, 6), I would like to assign different weights for each time series across models. Below is a sample loadData data frame illustrating the intended structure:

loadData <- expand.grid(
series = c("M1", "M2", "M750", "M1000"),
model = c("glmnet", "xgboost")
)
loadData$weight <- c(0.4, 0.5, 0.6, 0.9, 0.6, 0.5, 0.4, 0.1)
loadData

For example:
• series = M1, model = glmnet, weight = 0.4
• series = M1, model = xgboost, weight = 0.6

Could you advise how to implement this custom weighting setup within the ensemble_weighted function?

Thank you very much for your guidance.