Performing literature review using text mining, Part III: Summarizing articles using TextRank

Owing to the popularized utilization of academic search engines, such as Google Scholar or Microsoft Academic, the information that a researcher can conveniently access is unlimited. Especially in the domain of supply chain and transportation, where the academic publications are text-heavy and algorithm-driven, it is virtually impossible for one to follow the large amount of developments that is being created on a day-to-day basis. This poses a major barrier during literature review. To that end, the conventional literature-review methodology needs to be revised; text mining has strong potential in this aspect. In the earlier parts of this series of papers, technology infrastructure (Part I) and abbreviation extraction (Part II) have been studied. Both of those analyses focus on analyzing the words in a given set of documents. In Part III, moving from words to semantics, the task of automatic summarization of research papers is explored. More specifically, the TextRank algorithm is used to extract top N most important sentences from a paper, which could significantly boost the efficiency in scanning documents during literature review.