Semantic data integration for forest observatory applications

Hamed, Naeima ORCID: https://orcid.org/0000-0002-2998-5056 2024. Semantic data integration for forest observatory applications. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Thesis) - Accepted Post-Print Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (12MB) | Preview
	PDF (Cardiff University Electronic Publication Form) Restricted to Repository staff only Download (160kB)

Abstract

The populations of endangered species, such as African and Asian elephants, are declining due to habitat loss, fragmentation, and poaching. Driven by the ivory trade, poaching involves the unlawful killing of animals, posing a significant threat to elephant populations. Due to funding shortages in wildlife conservation, analysing research data has emerged as a cost-effective solution for decision-making in protecting wildlife species. Wildlife research data are typically collected on a project-specific basis, leading to the creation of data silos. Addressing key conservation questions often requires unified access to diverse data sources. For example, accessing elephants’ tracked movements alongside environmental data during the dry or wet season can help predict whether they are heading towards locations that expose them to poachers. This research introduces a novel approach that employs semantic web technologies to integrate heterogeneous wildlife data from the forests of Sabah in Malaysian Borneo. A review of Open Data Observatories and their data management methods identified the Semantic Web as an effective approach to breaking wildlife research data silos. Consequently, the Forest Observatory Ontology (FOO) was developed to standardise sensor-monitored wildlife data for integration. FOO was populated with four heterogeneous wildlife datasets to construct knowledge graphs. Predictive models derived from these knowledge graphs were used to predict elephants’ geo-locations and poaching likelihood, providing a proactive tool for conservationists. To extend the research, the generalisation of the methodology to different domains was explored by developing and populating another ontology for Internet of Things (IoT) data marketplaces, enabling on-demand data purchasing. This doctoral research contributes to wildlife data management by analysing Open Data Observatories to identify optimal approaches for integrating data. It develops the Forest Observatory Ontology (FOO) and its associated knowledge graphs to standardise and unify wildlife data generated by sensors. Using the constructed knowledge graphs, the research creates predictive models for poaching through deep learning and semantic reasoning.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Computer Science & Informatics
Date of First Compliant Deposit:	31 January 2025
Date of Acceptance:	30 January 2025
Last Modified:	03 Feb 2025 15:05
URI:	https://orca.cardiff.ac.uk/id/eprint/175791

Actions (repository staff only)

Edit Item