A comparative study of automated reviewer assignment methods

Abstract

The reviewer assignment problem is the problem of determining suitable reviewers for papers submitted to journals or conferences. Automated solutions to this problem have used standard information retrieval methods such as the vector space model and latent semantic indexing. In this work we introduce two new methods. One method assigns reviewers using compression approximated information distance. This method approximates the Kolmogorov complexity of papers using their size when compressed by a compression program, and then approximates the relatedness of the papers using an information distance equation. This method performs better than standard information retrieval methods. The second method assigns reviewers using Google desktop a more advanced information retrieval system. The method searches for key terms from a paper needing reviewers in a set of papers written by possible reviewers and uses the search results as votes for reviewers. This method is relatively simple and is very effective for assigning reviewers.