Semantic Document Networks to Support Concept Retrieval

Abstract

There are many unstructured documents created in many disciplines which need to be (pre-) processed in one way or another for further integration and use in IT systems. The predominance of the Internet and large corporate databases implies that there are large volumes of documents that need to be analysed and searched to retrieve information; particularly within the fields of machine translation, text analysis, semantic mining, information extraction and retrieval. We explicate a framework based on concept-based indexing that supports the analysis, storage, and retrieval of documents. Natural-language reduction is used to calculate semantic cores for concept-based indexing of stored concepts found within documents. The processed documents are stored within a semantic network enabling effective analysis of core concepts within documents and rapid retrieval of specific ideas from multiple documents based on provided concepts.