Natural Language Question Generation from Connected Open Data: A Study of Possibilities

Abstract

Accelerated growth of Linked Open Data has been observed. There are several motives for this. Some situations are the generation of these bases from texts, another is the amount of open data generated from information systems. Their growth results in a large volume of information available. This fact brings the possibility of large-scale use in question-and-answer systems. Which have a general function that is based on a structure of information to generate phrases and to conference of produced answers. From literature studies, we foresee an opportunity of bigger use, in several applications. It is possible to generate natural language phrases from linked open databases, in Portuguese. In addition, we identified challenges to effective applicate linked open data resources for this purpose effectively. In this way, this work aims to see which aspects of the structure of open linked databases could be used as support in generating natural language questions.