On-line algorithm to predict nearly as well as the best pruning of a decision tree
pp. 296-306
in: Setsuo Arikawa, Ayumi Shinohara (eds), Progress in discovery science, Berlin, Springer, 2002Abstract
We review underlying mechanisms of the multiplicative weight-update prediction algorithms which somehow combine experts' predictions to obtain its own prediction that is almost as good as the best expert's prediction. Looking into the mechanisms we show how such an algorithm with the experts arranged on one layer can be naturally generalized to the one with the experts laid on nodes of trees. Consequently we give an on-line prediction algorithm that, when given a decision tree, produces predictions not much worse than the predictions made by the best pruning of the given decision tree.