Generalization Bound for Multi-Classification with Push
Jin Luo, Yongguang Chen, Xuejun Zhou · 2010
Solving multi-classification problems has been improved by overcoming the limit of conventional statistical methods supported by development of artificial intelligence methods. The derived bound provides a means to evaluate clustering solutions in terms of the generalization power of a built-on classifier. For classification based on a single feature the bound serves to find a globally optimal classification rule. Comparison of the generalization power of individual features can then be used for feature ranking. In this paper we take multi-classification as push the sample on the top of the list, to different class, and derive a generalization bound for multi-classification by using covering number to provide a specific type of conclusion.