| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 100 === In the early response phase after the disaster happens, most responders need to find integrated and relevant information to make decision. In this case, finding suitable information in the open crowdsourcing environments is a complex task, since it involves many actors and a large amount of unstructured and heterogeneous spatial data. While quite significant progress on providing system and standardizing syntax heterogeneity of data has been made, semantic issues are still insufficiently addressed. Using ontological approach to represent information in disaster and emergency management can resolve this semantic heterogeneity problem. However, to the best of our knowledge, there is no formal vocabulary or ontology in existence that specifically allow victims to describe incident information in their nature language.
In this thesis, we build a domain ontology model for disaster and emergency information management which is able to capture descriptive spatial information about incidents and reasons them to get more valuable information that mostly needed by disaster responders. On the other hand, we also consider the possibility of integrating our ontology model with the other systems. In order to aim for these two objectives, we propose a methodology to implement hybrid ontological architecture by engaging SUMO (Suggested Upper Merged Ontology); we later complement it with knowledge-based representation that is capable of processing information depending on its content. Building on the result of our proposed ontology implementation, we create several experimental scenarios for information retrieval using our system.
|
|---|
