Searching and Ranking Documents based on Semantic Relationships

Abstract

Just as the link structure of the web is a critical component in today's web search, complex relationships (i.e., the different ways the dots are connected) will be an important component in tomorrow's web search technologies. In this paper, I summarize my research on answering the question of: How we can exploit semantic relationships of named-entities to improve relevance in search and ranking of documents? The intuition of my approach is to first analyze the relationships of namedentities with respect to a query. Second, relevance weights, which are assigned by human experts, can then be used to guarantee results within a relevance threshold. These relevance measures can be applied both for searching and ranking of documents.