Information extraction (IE) is the task of automatically extracting structured information from unstructured and semi-structured machinereadable document. In this paper, we propose a new paradigm for information extraction. In this extraction framework, intermediate output of each text processing component is stored so that only the improved component has to be deployed to the entire corpus. Extraction is then performed on both the previously processed data from the unchanged components as well as the updated data generated by the improved component. Performing such kind of incremental extraction can result in a tremendous reduction of processing time. To realize this new information extraction framework, we propose to choose database management systems over filebased storage systems to address the dynamic extraction needs. To demonstrate the feasibility of incremental extraction approach, experiments are performed to highlight two important aspects of an information extraction system: efficiency and quality of extraction results.
Keywords
Text mining, query languages, information storage and retrieval