Invasion Biology WikiProject Scientific Papers: Text Data Mining and LLM-based Information Extraction of Species, Locations, Habitats, and Ecosystems

<p>This dataset contains the abstract and full-text for publication DOIs from the Invasion Biology WikiProject (DOI:&nbsp;<a href="https://www.doi.org/10.5281/zenodo.12518036">10.5281/zenodo.12518036</a>). The data was retrieved using the <a href="https://ask.orkg.org/">ask.orkg.org</a> <a href="https://api.ask.orkg.org/docs#tag/Semantic-Neural-Search/operation/explore_documents_index_explore_get">API</a>. For the <a href="https://github.com/jd-coderepos/invasion-biology-IE/blob/main/scripts/ask-doi-list-fulltext-search.py">script</a> used to obtain the data, refer to the accompanying GitHub repository: <a href="https://github.com/jd-coderepos/invasion-biology-IE/" target="_blank" rel="noopener">https://github.com/jd-coderepos/invasion-biology-IE/</a>.</p> <p>The resulting CSV file includes the following fields: <code>"ASK ID"</code>, <code>"DOI"</code>, <code>"Title"</code>, <code>"Abstract"</code>, and <code>"Full-text"</code>.</p> <p>Of the 49,438 queried DOIs, the ASK database provided:</p> <ul> <li><strong>Total DOIs processed:</strong> 12,636</li> <li><strong>DOIs with neither abstract nor full-text:</strong> 36 (abstract token count was less than 10)</li> <li><strong>DOIs with abstracts but no full-text:</strong> 12,636</li> <li><strong>DOIs with both abstract and full-text:</strong> 2,834</li> </ul> <p>The second part of the dataset contains structured information extracted from the publications using the GPT-4o Large Language Model. This structured data is included in the zipped folder <code>structured-publications.zip</code>.</p> <p>The accompanying GitHub repository provides access to the code and scripts used at various stages of the information extraction (IE) process.</p> <p><strong>Theme of the Study:</strong><br>"Mining for Species, Locations, Habitats, and Ecosystems from Scientific Papers in Invasion Biology: A Large-Scale Exploratory Study with Large Language Models."</p>