Proceedings of the 4 th International Conference on Computing and Informatics, ICOCI 2013 28-30 August, 2013 Sarawak, Malaysia. Universiti Utara Malaysia (http://www.uum.edu.my ) Paper No. 036 371 A COMPUTATIONAL METHOD TO ANSWER SEMANTIC QUESTIONS WITH CONSTRAINTS Nurfadhlina Mohd Sharef 1 , Shahrul Azman Mohd Noah 2 , Rabiah Abdul Kadir 1 , and Azreen Azman 2 1 Intelligent Systems Research Group, Faculty of Computer Science and Information Technology, University of Putra Malaysia, Malaysia, fadhlina@fsktm.upm.edu.my , rabiah@fsktm.upm.edu.my, azreen@fsktm.upm.edu.my 2 Knowledge Technology Group, Centre of Artificial Intelligence, Faculty of Technology and Information Science, National University of Malaysia, samn@ftsm.ukm.my ABSTRACT. Semantic question answering is a discipline that allows the development of smart systems through its reasoning and natural language understanding capability. Knowledge is stored based on representation format such as ontology. However, querying these knowledge requires understanding of the knowledge structure and demands the user to be equipped with the formal query construction skill. This will definitely hinder the development of this area. Therefore this paper addresses the translation model of constraints-typed questions. We focus on the components that relate to constraints-typed question answering and propose a computational model for the translation. These figures indicate promising development in this problem and should encourage more alternative methods in the same direction. Keywords: Semantic, Question Answering, SPARQL, Natural Language INTRODUCTION Semantic question answering (SQA) over the steadily growing amount of semantic data opens possibilities not conceivable before; deep and accurate answering, compared to keyword based matching adopted by information retrieval approach. Through the semantic represented knowledge, reasoning is allowable by connecting and making sense of the content of the knowledge. However, due to the structured format, the knowledge cannot be benefited by the novice user without mastering the query language thus natural language (NL) question is the best media. On the other hand, the question may require composed information from several sources. Furthermore, variation of question complexity demands different execution strategies and thus traditional federated query approach is unsuitable. Although sources mapping can be performed, this is typically with low automation thus demands high human labor. Traditional information retrieval approach is insufficient to solve this problem because it cannot exploit the internal structure of the data. Besides, query federation strategy will also fail because it typically distribute the questions into the resources and integrate the answer. This will return very low accuracy because (i) automatic identification of the source to utilize is challenging, (ii) the result may not be suitable for straightforward integration. Therefore, formal query (i.e, SPARQL) needs to be constructed to retrieve the data from the semantic sources.