Using semantic and structural similarities for indexing and searching scientific papers

Finding relevant scientific documents from a huge set of academic papers is a challenging task and with the tremendous growth in electronic publication, locating the most relevant and related scientific documents when going through a new research paper is becoming even more challenging. In this paper, we present a new way of indexing and searching the scientific documents to assist researchers in finding relevant documents when coming across a new research document. In particular, we explored how DT-Tree (DocumentTerm-Tree) - a new structure for the representation of scientific documents - can be used to create an index of scientific documents. We used MVP-Tree to create index using DT-Tree representation of the documents. We then performed search experiments, using new scientific documents as queries, to show that relevant documents are retrieved when DT-Tree structures are used to create MVP-Tree.