You have cleaned your data and removed all correlating features. You have also visualized your dataset and know the class labels are separable. You have also tuned your hyper-parameters. That’s great, but why isn’t your model performing well?
Did you try stacking your models? Traditionally, we have modeled our data with a single algorithm. That might be a Logistic Regression, Gaussian Naive Bayes, or XGBoost.
Image by Author: Traditional ML Model
Here’s what an ensemble stacking model does:
Image by Author: Stacked ML Model
An algorithm that is used to combine the base estimators is called the meta learner. We can determine how we want this algorithm to respond to different predictions from other models(classifiers in this case). It can be:
#model-stacking #machine-learning #classification #ensemble-learning #deep learning