Author Disambiguation by mining the coauthor graph

Nowadays, we rely mostly on search engines when surveying academic papers. Many mainstream search engines provide searches specific to academic articles such as Google Scholar, Microsoft Academic Search … etc. Many conferences also host their online paper archive. One of the challenging problems to these systems is author name disambiguation. An author may represent their name differently in different locations. Because the name is one of the most keyword used to search papers, it became crucial for search engines to recognize different names from the same person to generate accurate and complete results. Through author name ambiguity may be caused by many reasons, one of the most common reasons that lead to this author identity ambiguity is different name representations. Solving this problem generally require much supplementary information about authors that may not be available. In this paper, we purpose a system that tries to resolve the author name ambiguity issue caused by the different way the author write his/her name, the only information it requires is author’s name and co-author relationship. The experiment shows the effectiveness of our system compared to some traditional methods.