The problem with ranking ensembles based on training or validation performance.

Ulf Johansson Tuve Löfström Henrik Boström

Published in: IJCNN (2008)

Keyphrases

diversity measures
test set
ranking algorithm
decision trees
training set
supervised learning
loss function
feature ranking
training process
ranking functions
web search
data sets
learning to rank
link analysis
feedforward neural networks
training data
genetic algorithm