News Hunter: a semantic news aggregator

Abstract

This thesis presents a research project where a semantic news aggregator system called News Hunter was developed and evaluated. News Hunter was a collaboration between UiB and Wolftech that aimed at merging semantic web technologies with newsroom systems. The goal was to provide those working in newsrooms with quick and easy access to relevant background information during their research and writing process. To achieve this goal we built a system that could extract named entities and keywords from incoming text, and then store and use these to gather data from other resources. News Hunter was capable of pre- senting journalists with up-to-date background information on incoming news messages, as well as live-updated information when writing new stories. Journalists and domain experts tested the system, and answered questions about the usefulness of each feature. Some fea- tures were tested with F1-scoring and others by comparing it to publicly available systems with similar functionality. The accuracy of the system and the results from user evaluation give encouragement for further development of semantic newsroom systems and additional research into the field.