Measures in Collection Ranking EvaluationZhihong

As a technique to hunt information on the Internet, collection location has received more attention. Several approaches have been proposed to solve this problem. All these approaches adopt the same procedure: ranking the collections and returning the top-ranks. But these approaches deene diierent measures to evaluate collection ranking and the measures have signiicant weaknesses. In this paper, we survey the measures used in current research and propose a new pair of measures that are based on the concepts of precision and recall. The new measures overcome the problems found in the current measures.