Ontology based Spreading Activation for NLP related Scenarios

erschienen November 2011 in: Lissabon, Portugal SEMAPRO 2011

To handle the flood of information in the modern world new technologies are needed. One problem is the handling and filtering of information itself. Semantic technologies have been named to possess the potential to at least facilitate this problem. Another difficulty is the representation of information to humans. Different algorithms and user interface concepts have been created allowing the access on a very specific type and structure of information. However the most common and natural way for humans is to use natural language. Natural Language Processing tries to analyze the syntax and semantics of language, but often delivers unsatisfying results because of the many phenomena (e.g., ambiguity) humans use while communicating. We therefore currently develop an approach, which allows us to analyze the semantic content of natural language text based on an ontology. In this paper we present a spreading activation based algorithm, which not only helps identify the correct semantic concepts for a natural language text, but also partially solves other phenomena of natural language.