Breiman, Friedman, Gordon and Stone recognized that tree classifiers would be very valuable to practicing statisticians. Their cart algorithm became very popular indeed. Designing tree-based classifiers, however, has its pitfalls. It is easy to make them too simple or too complicated so that Bayes risk consistency is compromised. In this talk, we explore the relationship between algorithmic complexity of tree-based methods and performance. © 2010 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Devroye, L. (2010). Classification and trees. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6218 LNCS, pp. 40–44). https://doi.org/10.1007/978-3-642-14980-1_3
Mendeley helps you to discover research relevant for your work.