Using Lotkaian Informetrics for Ranking in Digital Libraries

Abstract

The purpose of this paper is to propose the use of models, theories and laws in bibliometrics and scientometrics to enhance information retrieval processes, especially ranking. A common pattern in many man-made data sets is Lotka's Law which follows the well-known power-law distributions. These informetric distributions can be used to give an alternative order to large and scattered result sets and can be applied as a new ranking mechanism. The polyrepresentation of information in Digital Library systems is used to enhance the retrieval quality, to overcome the drawbacks of the typical term-based ranking approaches and to enable users to explore retrieved document sets from a different perspective.