Mexpert: An Algorithm for Finding Cross-disciplinary Experts Using Data Mining Techniques

Abstract

Multidisciplinary research is the practice of bringing together experts from different disciplines or fields of study to work collaboratively on a common problem, project, or research question. This study proposes an expert finding algorithm, namely Mexpert, to bridge the gap between experts using data mining techniques. Our method has three steps. The first process is data preparation, which involves extracting information from research papers. The first process includes collecting data, extracting keywords, and discovering topics. The second process involves creating a model to group experts with similar knowledge and expertise. The last process involves analyzing the expert profiles using their experience in terms of h-index, average number of publications and total citations, followed by profile ranking. The corpus contains 17,679 papers obtained from SCOPUS. The experimental results reveal that there are four clusters out of 7 fields. We analyzed each cluster obtained from Mexpert and found that most cluster members published multidisciplinary research papers together. These results suggested that our approach can be applied to find a group of experts with different expertise.