Answering Engine for sports statistics: Development of an ontology and a knowledge base

Abstract

Focus of this study is on development of a domain-specific ontology and a knowledge base of facts for a Question Answering system. This QA system accepts natural language questions about statistics of Formula One and transforms them in to formal queries using natural language processing techniques and the designed ontology. It then executes the queries against the knowledge base to return exact answers.

During the design process of the ontology, regular standards and regulations have been utilized, and the required data for implementing the ontology have been collected from a large-scale and reliable data source. Semantic technologies have been used to transform data to structured and machine-readable formats and a graph knowledge base is used for storage and retrieval of the structured data through formal queries.

The evaluation results show that the knowledge base covers lots of correct and relevant information about main entities in the given domain. The designed ontology has required potential to answer many statistical questions that it was designed for and the QA system based on this ontology can provide correct answers to easy questions about statistics of Formula One. The limitation of the ontology is that it cannot provide the QA system with the necessary knowledge to answer complex queries about statistics of Formula One.