Indexing of authors according to their domain of expertise

Abstract

Measuring the impact and productivity of an author is an important, yet a challenging task. Most of the existing methods for ranking or indexing of authors are based on simple parameters such as publication counts, citation counts and their combinations. These methods are topic independent, hence ignoring the intra-field differences. This study introduces a specific method for indexing of researchers to measure their productivity in a given field of interest, believing that an author can be interested in more than one fields and can have different level of expertise in all these fields. This paper proposes Domain Specific Index (DSI), a novel method for indexing of authors with respect to their fields of interest. Latent Dirichlet Allocation (LDA) is applied to capture the latent topics within text corpora. DSI calculates the standing of an author in all topics of his or her interest by considering topic based citations instead of using overall citations like traditional methods. The citations received by a multi-authored paper are divided among all its co-authors on the basis of their topic probability in that particular field. Results show that instead of giving credit of received citations equally to all co-authors of a paper, if a weight is given with respect to their level of interest in that field, more specific authors in that field will be ranked as top authors.