SUSAX: Context-specific searching in XML documents using sequence alignment techniques

Keyword searching while very successful in narrowing down the contents of the Web to the pertaining subset of information, has two primary drawbacks. First, the accuracy of the search is closely coupled with the choice of keywords. Second, keywords are limited in their expressibility. In particular, they fail to adequately capture the contextual information implicit in most searches done by users. In this paper we present an approach to efficiently address these drawbacks of keyword searching over XML documents. In particular, we present SUSAX a system for approximate contextual querying over XML documents wherein queries are represented as simple XPaths. A key contribution of our work is the novel algorithm used to match the XPath-like query with similar paths in the repository. The algorithm is based on sequence alignment algorithms prevalent in life sciences domain for discovering the similarity between genome and protein sequences. In this paper, we show an adaptation of the sequence alignment algorithm for now discovering and cataloging the similarity between two paths.