Cross-Lingual Ontology Matching and Enrichment

Abstract

The Semantic Web has emerged as a foundational paradigm for representing and interlinking data in a semantically meaningful manner. Central to this paradigm are ontologies, pivotal for encapsulating spatial knowledge and fostering a standardized comprehension of information. However, the multilingual nature of global communication presents challenges to the seamless interoperability and effective utilization of ontologies, such as semantic ambiguity across languages and the limited availability of bilingual dictionaries. <br /> This thesis aims to tackle these challenges by developing innovative approaches to utilizing multilingual data in order to enrich ontologies across different languages. The ultimate goal is to facilitate semantic interoperability and the exchange of knowledge across ontologies in a multilingual context. Key steps towards this goal have been taken by proposing approaches for cross-lingual ontology matching and enrichment, accompanied by the development of a comprehensive metrics suite to assess the quality of the enrichment process. <br /> In order to achieve this goal, a cross-lingual ontology matching approach is introduced to align ontologies cohesively before the enrichment process. This approach provides a robust framework for evaluating various similarity measures across languages, empowering researchers to select the most suitable measures for their endeavors. <br /> Cross-lingual ontology enrichment is identified as a cornerstone, enabling the representation of domain-specific knowledge written in different languages through enriching ontologies with additional linguistic and semantic content. We proposed approaches for enriching ontologies by leveraging information from unstructured multilingual text corpora and diverse ontological resources, producing multilingual ontologies. These approaches promote the efficient integration and sharing of knowledge across systems that operate in multilingual contexts. <br /> One of the main obstacles to improving the quality of the enrichment process is ensuring the accuracy and relevance of additional information. Therefore, a thorough analysis of factors influencing the enrichment process has been conducted, resulting in the innovation of a comprehensive quality metrics suite. <br /> Through empirical evaluations and real-world use cases, the efficacy and applicability of the proposed approaches are demonstrated, highlighting their potential to facilitate seamless cross-lingual communication, enhance knowledge sharing, promote semantic interoperability, and promote cultural diversity in the Semantic Web.