I read some stuff about majority vote and greedy action in ensembling, however, they kind of sound similar, but also different. What is the real difference between those two?
Thanks for your help!
Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.Sign up to join this community
Ensemble learning basically means combining several base models to produce one optimal solution.
And from my understanding, usually in Majority vote, we train the models with all the data and wait for all the models to result in a prediction and at the end we choose the major result (result that appeared more than any other) as the final output.
But in Greedy way, we usually iterate on the data, run prediction on a subset of the data, pick the best result of this subset and move to the next, and at the end we only consider the best results of each iteration. So we don't look at the big picture here but only a collection of subsets. Not the Global best result but Local best results.
Hope this helps.