Automatic Summarization of Scientific Articles from Biomedical Domain

Automatic text summarization is a well-known topic in the field of natural language processing. Text summarization is a process of finding the gist of some text document. Automatic text summarization does that work with computer without any human help. With the increase of information, it becomes difficult to efficiently utilize them. Leveraging the methods of natural language processing, automatic text summarization can efficiently summarize a large collection of information. In this work, we model extractive text summarization using WordRank, a modified version of TextRank. We showed that combining WordRank with TextRank improves the quality of the summary marginally. We experimented with different combinations of the two algorithms to find a better summary of the PubMed articles. Our proposed method works significantly better than previous two methods known as TF-IDF and paragraph extraction.