Extracting semantic annotations and their correlation with document components

Abstract

Digital repositories can preserve terabytes of information in the form of digital documents. Searching these digital documents requires time and computing resources. Techniques are required to efficiently process these digital repositories. Metadata and semantic annotations can augment the overall search process and provide a foundation to build intelligent applications by using the documents in the repository. In this paper, we are proposing an approach for generation of context aware metadata to enhance search for the scientific publications and also prove the impact of compound words on semantic metadata. Major contribution of our work is to correlate the extracted semantic annotations with the document components. This allows, for example, searching a document centered around a scientific claim by differentiating between author's claims and statements about related systems mentioned in different document components. The approach utilizes the syntactic and semantic measures to increase the quality of the extracted semantic annotations and to bring improvements in precision of search results.