Similarity-based Complex Publication Network Analytics for Recommending Potential Collaborations

As communities of researchers continue to become quite large and to grow incessantly, collaboration among researchers can be conducive to greater research productivity. Nevertheless, it is difficult for a researcher to find suitable collaborators from all researchers around the world. In this paper, we have used bibliographic DBLP data to extract information of a researcher and to discover the relationship between the co-authors and between authors and conferences. We evaluated some of the similarity measures and developed an innovative random walk model to find potential co-authors for a given researcher. These measures were then used to design a best model to recommend co-authors. We have also applied an HITS algorithm and proposed a ranking algorithm to rank researchers and conferences with the intent of recommending authors or conferences.