Hi everyone,
I have one question. I’m attempting to implement Weka Stacking 3.7 in my workflow, but I’ve encountered a challenge. The Stacking 3.7 learner already includes an integrated cross-validation approach. Due to this, I had the idea of using an x-partitioner prior to this node (number of validations = 1) to divide the data into training and testing sets. However, as the Stacking node itself employs its own cross-validation, I’m a bit unsure about the most effective way to perform this data division since the number of validations = 1 is not possible to apply, only greater than 1. If I don’t split the data before this node, I’ll essentially be providing the entire dataset to the Weka predictor as well.
Could someone help me?