[翻译][Paper][WWW'10]Classification-Enhanced Ranking

ABSTRACT
Many have speculated that classifying web pages can improve a search engine's ranking of results. Intuitively results should be more relevant when they match the class of
 a query. We present a simple framework for classi cation-enhanced ranking that uses clicks in combination with the classi cation of web pages to derive a class distribution for the query. We then go on to de ne a variety of features that capture the match between the class distributions of a web page and a query, the ambiguity of a query, and the coverage of a retrieved result relative to a query's set of classes. Experimental results demonstrate that a ranker learned with these features signi cantly improves ranking over a competitive baseline. Furthermore, our methodology is agnostic with respect to the classi cation space and can be used to derive query classes for a variety of di erent taxonomies.

摘要：

许多推断认为网页分类可以提高搜索引擎的排序结果。当搜索结果匹配到一个query的类型时我们很直观地认为搜索结果与query相关。我们将介绍一种简单的网页分类框架用于加强分类式排序，它使用了点击分类网页，从而获得一条query的类型分布。然后我们继续定义了多种匹配类型分布与query、query的歧义、query的类型集合关联的检索结果覆盖率的特征。实验结果证明排序算法通过学习这些特征可显著提高排序超越了原本的竞争基线。此外，我们的方法论对分类空间时不可知的，并且可以用于获取遵守多种分类标准的query分类结果。