New Geology Text Mining Method Enhances Automated Extraction of Geological Information

A study led by Prof. ZHANG Nannan from the Xinjiang Institute of Ecology and Geography (XIEG) of the Chinese Academy of Sciences has introduced an innovative geological knowledge-constrained method for extracting entities and relationships from textual data. Their work was published in Computers & Geosciences.

The team developed a specialized extraction model that incorporates geological knowledge constraints to identify triples—comprising entities and their relationships—from geological texts. They designed a geological schema specifically tailored to granitic pegmatite-type lithium deposits, which includes 22 entity types, 16 relation types, and 184 knowledge rules. Using a dataset constructed from 68 published research articles and four mineral exploration reports, the researchers integrated this schema into their information extraction model as a constraint mechanism.

Their findings showed that embedding the geological schema not only increased computational efficiency but also improved the accuracy of the extraction process. The study demonstrates the effectiveness and practicality of the proposed model for geological information retrieval.

"This framework offers a novel technical approach to geological information extraction," said TAO Jintao, first author of the study.

This study provides a clear roadmap for the automatic extraction of entity-relation triples from geological texts, marking a significant advancement in geological text mining and knowledge discovery workflows.

Read the full article: https://doi.org/10.1016/j.cageo.2025.105920

Contact

LONG Huaping

Xinjiang Institute of Ecology and Geography

E-mail: longhp@ms.xjb.ac.cn

Web: http://english.egi.cas.cn