A Question Answering System Based on Conceptual Graph Formalism

This paper proposes a new text-based question answering system. It models knowledge in documents and questions with conceptual graph formalism (CGF). To prepare knowledge to be modeled, natural language processing is applied to the text using OpenNLP, and then syntactic and semantic information is realized using VerbNet and WordNet. We handle different types of questions, especially questions constructed with wh- pronouns, and questions constructed with ¿how¿, and we propose a model to represent them in CGF so that their target is realized and marked. Each question´s type has different conceptual graph (CG) representations; thus, for each question, many CGs are generated using formulas. Some of these formulas are introduced here. The projection operator is used to compare a question´s CG to a sentence´s CG, and then the exact answer is extracted from the part of the sentence´s CG that has been projected under the question target´s concept.