Medtree: A Search Engine for Medical Professionals

Overview & Synopsis

Users struggle with keyword based search engines like Google or Bing because queries can have multiple interpretations and search engines fail to understand the context in which the user is looking for information. This failure leads to search results that are either inappropriate or contextually irrelevant. In this paper we describe algorithms which, utilizing information about the user's context, scrape the web and process/filter candidate sites that could be used to create a customized context specific search engine for the user. We used the algorithm to create Medtree, a customized medical search engine for doctors at Stanford Hosptial. In evaluations we demonstrate Medtree's superiority to Google for medical queries.