Adaptive information extraction from unstructured documents

Abstract

The authors present a novel adaptive framework that enables efficient development of applications demanding complex document analysis. In processing natural language documents the task is to transform them into application specific structured form. Such transformation has to be designed taking into account various abstraction levels and granularity of the processing and the multitude of possibly related requests driving the application. The proposed solution is based on the adaptively planned and executed network of information processing modules. The paper presents an overview of the framework, with the focus on the adaptive mechanism. An illustrative pilot application is also provided.

Intelligent information extraction from scholarly document databases

“…Natural language processing (NLP) is a constant reference in most publications (Hassan and Le 2020). Sometimes their proposals ask for structured documents and, when not, they need to transform documents into structured data (Dezsenyi et al 2007;Oro and Ruffolo 2008). Other times they need to convert the original PDF files into HTML and text format files to be able to proceed ( The documents analyzed propose algorithmbased systems and agents with rules to query document databases, although it is common to find unsolved problems when there are heterogeneous data sources (Seng and Lai 2010).…”

“…x x x x x x x x x x Journal Article Dezsenyi, C., Dobrowiecki, T. P., and Meszaros, T. 2007 Adaptive information extraction from unstructured documents Sistema. Transformación documento a estructurado.…”

Untitled

“…Natural language processing (NLP) is a constant reference in most publications (Hassan and Le 2020). Sometimes their proposals ask for structured documents and, when not, they need to transform documents into structured data (Dezsenyi et al 2007;Oro and Ruffolo 2008). Other times they need to convert the original PDF files into HTML and text format files to be able to proceed (Hassan and Baumgartner 2005a;Rizvi et al 2018;Seng and Lai 2010).…”