Abstract:
The Internet and World Wide Web have been the driving technologies for this information era. The vast amount of information, both structured and unstructured, that is published on the web becomes a pool of knowledge for web users. In this research, we propose a method to provide linkage between a natural language text (i.e. unstructured data) and the linked open data in DBpedia (i.e. structured data) in order to provide semantic enhancement of the text. Specifically, we present the Knowledge Enhancement of Text API (KnET) that, based on DBpedia Spotlight, can identify entities in a text document which have corresponding DBpedia resources. The KnET API then discovers the semantic relations between those DBpedia resources. The research also presents an example use case of this API to build a visual knowledge enhancement web application that can visualize the relations between those DBpedia resources in the form of a graph. Such a knowledge graph can complement the text by giving information that is additional to that found in the text. In an experiment on students in grades 5 and 6 which are assigned to read English passages, knowledge of the students about the topics in the passages is statistically significantly improved after using the visual knowledge enhancement application. The application also scores high in terms of user satisfaction.