Besides web resources that are visible to traditional search engines such as Google or Bing, there are a large portion of the web that is invisible to them. For example, the Purdue Library has access to more than 400 databases. Most of them contain documents that are not reachable by traditional web crawlers due to security, technical limitations or copyright agreement. The only way to search and retrieve documents from those databases is through their service interfaces, which implement different methods of indexing and searching. Given a large number of databases available, it is difficult for users to select the best databases for their searches. In this poster, we present an intelligent search system for Purdue Library. This system is able to analyze users' queries, select the most relevant databases, retrieve documents from those databases and represent the merging result to users. This poster was presented at the 2009 IT Summit at Purdue University.
Federated search, information retreival, electronic resource selection, library science, computer science
Date of this Version