On the Practice of Branching Program Boosting
نویسندگان
چکیده
Branching programs are a generalization of decision trees. From the viewpoint of boosting theory the former appear to be exponentially more eecient. However, earlier experience demonstrates that such results do not necessarily translate to practical success. In this paper we develop a practical version of Mansour and McAllester's 13] algorithm for branching program boosting. We test the algorithm empirically with real-world and synthetic data. Branching programs attain the same prediction accuracy level as C4.5. Contrary to the implications of the boosting analysis, they are not signiicantly smaller than the corresponding decision trees. This further corroborates the earlier observations on the way in which boosting analyses bear practical signiicance.
منابع مشابه
Evaluation of the effect of dendritic branching on signal processing in hippocampus pyramidal cells
Since branching region of an active nerve fiber is an abrupt widening of the structure, two concepts emerge: first, the stimulating current must be sufficient to raise the outgrowing fibers above the thresh¬old, and secondly, the stimulating current will be divided in proportion to the characteristic admittance of the branches. On the other hand, blocking of the nerve impulse in this region is ...
متن کاملEvaluation of the effect of dendritic branching on signal processing in hippocampus pyramidal cells
Since branching region of an active nerve fiber is an abrupt widening of the structure, two concepts emerge: first, the stimulating current must be sufficient to raise the outgrowing fibers above the thresh¬old, and secondly, the stimulating current will be divided in proportion to the characteristic admittance of the branches. On the other hand, blocking of the nerve impulse in this region is ...
متن کاملBoosting Using Branching Programs
It is known that decision tree learning can be viewed as a form of boosting. Given a weak learning hypothesis one can show that the training error of a decision tree declines as |T| where |T| is the size of the decision tree and b is a constant determined by the weak learning hypothesis. Here we consider the case of decision DAGs—decision trees in which a given node can be shared by different b...
متن کاملInteractive Martingale Boosting
We present an approach and a system that explores the application of interactive machine learning to a branching program-based boosting algorithm—Martingale Boosting. Typically, its performance is based on the ability of a learner to meet a fixed objective and does not account for preferences (e.g., low FPs) arising from an underlying classification problem. We use user preferences gathered on ...
متن کاملBoosting with Multi-Way Branching in Decision Trees
It is known that decision tree learning can be viewed as a form of boosting. However, existing boosting theorems for decision tree learning allow only binary-branching trees and the generalization to multi-branching trees is not immediate. Practical decision tree algorithms, such as CART and C4.5, implement a trade-off between the number of branches and the improvement in tree quality as measur...
متن کامل