🤖 AI Summary
To address low accuracy and poor generalization in automatic geolocation extraction from news texts, this paper proposes the first retrieval-augmented generation (RAG) framework tailored for geocoding. Our method jointly leverages a geographic knowledge base and large language models (e.g., Llama-3, Phi-3) to perform end-to-end latitude–longitude generation. It enhances mapping accuracy via context-aware prompt engineering and a multi-source location disambiguation mechanism. Evaluated on three standard benchmarks, our approach significantly outperforms traditional rule-based systems and fine-tuned baselines, achieving an average 12.7% F1-score improvement. Ablation studies confirm synergistic gains between the retrieval and reasoning modules. To ensure reproducibility and practical utility, we fully open-source the code, datasets, and models—enabling applications such as epidemic surveillance and disaster response geocoding.
📝 Abstract
Geocoding involves automatic extraction of location coordinates of incidents reported in news articles, and can be used for epidemic intelligence or disaster management. This paper introduces Retrieval-Augmented Coordinate Capture Of Online News articles (RACCOON), an open-source geocoding approach that extracts geolocations from news articles. RACCOON uses a retrieval-augmented generation (RAG) approach where candidate locations and associated information are retrieved in the form of context from a location database, and a prompt containing the retrieved context, location mentions and news articles is fed to an LLM to generate the location coordinates. Our evaluation on three datasets, two underlying LLMs, three baselines and several ablation tests based on the components of RACCOON demonstrate the utility of RACCOON. To the best of our knowledge, RACCOON is the first RAG-based approach for geocoding using pre-trained LLMs.