Abstract

The interpretation and analysis of historical mining data pose significant challenges due to its heterogeneous and scattered nature across various archival sources. This data is crucial for identifying new sources of critical minerals, understanding past resource utilization, and aiding in future project development, necessitating spatial, temporal, and semantic integration for comprehensive domain insights. In this paper, we detail our methodology for constructing, modeling, and semantically enriching a knowledge graph (KG) centered on historical mining data. Leveraging a custom ontology and semantic web technologies, we transform digitized archival records into a temporally and spatially aware, semantically rich KG. The resulting KG facilitates advanced temporal and spatial analyses through SPARQL queries, and enhances semantic richness by linking to additional data on the web. We demonstrate the application of our KG in the nuanced analysis of historical mining data and the generation of grade and tonnage models for two critical minerals: nickel and zinc. Our evaluation highlights the KG’s effectiveness in spatial and temporal interpretation of mining data, underscores the strengths of our entity linking method with an open knowledge base, and details the performance analysis of query execution. We also make the resulting KG available as open linked data.

Metadata

publication

year

2024

publication date

2024

authors

Basel Shbita, Namrata Sharma, Binh Vu, Fandel Lin, Craig A Knoblock

link

https://ceur-ws.org/Vol-3743/paper1.pdf

resource_link

https://ceur-ws.org/Vol-3743/paper1.pdf