SkylineSearch: Semantic Ranking and Result Visualization for PubMed

摘要

Life sciences researchers perform scientific literature search as part of their daily activities. Many such searches are executed against PubMed, a central repository of life sciences articles, and often return hundreds, or even thousands, of results, pointing to the need for data exploration tools. In this demonstration we present Sky-lineSearch, a semantic ranking and result visualization system designed specifically for PubMed, and available to the scientific community at skyline.cs.columbia.edu. Our system leverages semantic annotations of articles with terms from the MeSH controlled vocabulary, and presents results as a two-dimensional skyline, plotting relevance against publication date. We demonstrate that SkylineSearch supports a richer data exploration experience than does the search functionality of PubMed, allowing users to find relevant references more easily. We also show that SkylineSearch executes queries and presents results in interactive time.