Events Discovery Assistant: A Semi-Supervised Spatio-Temporal and Semantic Model for Discovering First World War Events

摘要

The present paper proposes a dynamic semantic search interface for discovering First World War events. A semi-automated knowledge retrieval algorithm was designed to automatically extract historical events from unstructured scholar texts and define their ontology-based semantic relevance, as well as, their spatial and temporal information. The interface is called "Events Discovery Assistant" and is able to improvise with the continuously growing historical contents. Based on semantic web technologies, the interface allow users to query events related to First World War specific (ontology-based) topic(s) and explore their evolved spatial and temporal patterns.