Approaches to passage retrieval in full text information systems
SIGIR '93 Proceedings of the 16th annual international ACM SIGIR conference on Research and development in information retrieval
Pivoted document length normalization
SIGIR '96 Proceedings of the 19th annual international ACM SIGIR conference on Research and development in information retrieval
An application of plausible reasoning to information retrieval
SIGIR '96 Proceedings of the 19th annual international ACM SIGIR conference on Research and development in information retrieval
| Hi-index | 0.00 |
Information Retrieval (IR) systems are built with different goals in mind. Some IR systems target high precision that is to have more relevant documents on the first page of their results. Other systems may target high recall that is finding as many references as possible. In this paper we present a method of document representation called RDR to build XML retrieval engines with high specificity; that is finding more relevant documents that are mostly about the query topic. The Rich Document Representation (RDR) is a method of representing the content of a document with logical terms and statements. The conjecture is that since RDR is a better representation of the document content it will produce higher precision. In our implementation, we used the Vector Space model to compute the similarity between the XML elements and queries. Our experiments are conducted on INEX 2004 test collection. The results indicate that the use of richer features such as logical terms or statements for XML retrieval tends to produce more focused retrieval. Therefore it is a suitable document representation when users need only a few more specific references and are more interested in precision than recall.