Forging a Path to Understanding: Integrating and Analyzing War History Data through Knowledge Graph Technologies

In the last two decades, the world has witnessed a steady increase in conflicts, violence, regional disputes, and local wars, resulting in significant casualties and economic losses. The impact of war on people's lives is severe and far-reaching, and therefore, peace has always been a shared aspiration. The study of history is an essential tool for understanding the dynamics of conflicts and the factors that contribute to their emergence. Examining the causes of past wars can help us to identify imbalances and potential flashpoints, and develop strategies and targeted interventions to prevent the onset of future conflicts. Understanding the nature of war is a complex undertaking that requires the integration of a significant amount of data across diverse fields. Data in the war field is often dispersed across different institutions, states, and languages, making it difficult for researchers to access and analyze the comprehensive range of resources necessary to gain a comprehensive understanding of war. To achieve a more comprehensive understanding of war, researchers must integrate data from diverse sources, curating and analyzing data to identify patterns, trends, and causal relationships that can inform decisions. The field of digital humanities has emerged as a result of the rapid development of information technologies and interdisciplinary research. One of the notable advancements in this field is the development of knowledge graph technologies. Knowledge graphs capture, represent, and organize real-world knowledge in graphs, with entities and relations to express the implicit connections of objects. Knowledge graphs not only provide a large amount of data but also emphasize the contextual information of the data. By integrating heterogeneous data sources into a single knowledge graph, researchers can gain a more comprehensive understanding of the complex nature of wars, including their causes, outcomes, and effects. The primary objective of this thesis is to explore the use of knowledge graph technologies for integrating heterogeneous data related to wars and to provide theoretical analysis and practical methods for its application in related fields. To examine the feasibility of the proposed methods, the Second Sino-Japanese War has been chosen as a case study. This research aims to provide insights into the practical application of knowledge graph technologies in the field of war-related data integration. This thesis makes several contributions to the field of war data integration. It has developed a multilingual domain-specific ontology that provides a standardized representation of concepts related to war, thereby enabling the integration of diverse sources of data and facilitating interoperability among different systems. An annotation dataset has been created for training machine learning models in the task of event extraction from war-related texts. The dataset has been carefully curated, and its creation involved the manual annotation of a large number of documents by human experts. This thesis has also proposed a joint model for event extraction tasks that integrates different types of information. The model is designed to address the challenges of event extraction from unstructured war texts, and its effectiveness has been demonstrated through extensive experiments and evaluations. This thesis also analyzes the characteristics and bottlenecks of the Second Sino-Japanese War data, discusses the feasibility and applicability of using knowledge graph technologies in war-related data, and evaluates the reuse of existing ontologies on the data. These discussions provide important insights into the challenges and opportunities associated with working with war data and suggest directions for future research in this area. The findings of this research have the potential to advance our understanding of war-related events and their representation in natural language

Author dc:creator

• Wang, Qian

Language dc:language.iso

en_AU

Handle dc:identifier.uri

http://hdl.handle.net/1885/305663

OAI identifier oai:identifier

oai:openresearch-repository.anu.edu.au:1885/305663