Error coding and PaCT's
暂无分享，去创建一个
A new class of plug in classi cation techniques have recently been developed in the statistics literature. A plug in classi cation technique (PaCT) is a method that takes a standard classi er (such as LDA or nearest neighbors) and plugs it into an algorithm to produce a new classi er. The standard classi er is known as the Plug in Classi er (PiC). These methods often produce large improvements over using a single classi er. In this paper we investigate one of these methods and give some motivation for its success.
[1] Yoav Freund,et al. Experiments with a New Boosting Algorithm , 1996, ICML.
[2] Thomas G. Dietterich,et al. Solving Multiclass Learning Problems via Error-Correcting Output Codes , 1994, J. Artif. Intell. Res..
[3] Leo Breiman,et al. Bias, Variance , And Arcing Classifiers , 1996 .
[4] Gareth M. James,et al. Generalizations of the Bias/Variance Decomposition for Prediction Error , 1997 .