A Multiview Clustering Approach To Identify Out-of-Scope Submissions in Peer Review

Despite criticisms, peer review is still the only widely accepted method for research validation. The first stage in academic peer review begins at the editor's desk where one essential job of the editor is to identify and reject inappropriate out-of-scope submissions. Here in this work, we investigate if we could assist the editor in identifying potential out-of-scope submissions. We view a paper from multiple perspectives and devise a multiview clustering approach to group in-scope and out-of-scope articles. Our semisupervised approach requires less training data yet achieves high performance. Our initial investigation yields promising results and has the potential to reduce the first turn-around time for journal submissions.