Ontology-based Question Answering System for Telecommunications with Document Extraction & Classification Abilities

Reza Hejzai, Mahmood; Sadat Mirian, Maryam; Neshatian, Koorosh; Reza Ofoghi, Bahador; Darrudi, Ehsan

Ontology-based Question Answering System for Telecommunications with Document Extraction & Classification Abilities

Authors

Mahmood Reza Hejzai

Maryam Sadat Mirian

Koorosh Neshatian

Bahador Reza Ofoghi

Ehsan Darrudi

Abstract

In this paper we present our ontology-based question answering system, named TeLQAS, designed and prototyped for the domain of telecommunications. TeLQAS is composed of two relatively independent processes. In the online process, the system accepts English natural language questions and provides exact answers reasoned from the ontology along with relevant summarized paragraphs from text resources. In the offline process, the system uses a text categorization mechanism to automatically extracted and classify relevant documents from available resources such as web and/or internal text warehouse. The conducted experiments show the outstanding results of the system with the specific domain of Fiber Optics while the accuracy of the system is improved with a larger question set. The method is general and can be deployed easily to other specific domains creating the corresponding ontology in the new domain.

Keywords

Question answering system

Information Retrieval

ontology

text summarization

Classification