PatentsView-Evaluation

PatentsView-Evaluation是由杜克大学和美国研究所共同开发的Python包，专注于评估发明人名称消歧系统的性能。该数据集包含多个基准数据集和评估工具，旨在通过提供高质量的评估数据和改进评估标准来推动发明人名称消歧的研究。数据集内容包括美国专利商标局2015年消歧竞赛中使用的数据集、Azoulay的学术生命科学数据集等，特别为评估目的开发。创建过程中，数据集通过关联专利号和发明人提及名称来准备。该数据集主要应用于解决专利数据中发明人名称的消歧问题，提高专利数据的质量和可用性。

PatentsView-Evaluation is a Python package jointly developed by Duke University and the American Research Institutes, focusing on evaluating the performance of inventor name disambiguation systems. This dataset encompasses multiple benchmark datasets and evaluation tools, aiming to advance research on inventor name disambiguation by providing high-quality evaluation data and improving evaluation standards. Its content covers the datasets used in the 2015 United States Patent and Trademark Office (USPTO) disambiguation competition, Azoulay's academic life science dataset, and other relevant resources, all specially developed for evaluation purposes. The dataset was constructed by linking patent numbers with inventor mentions during its development. This dataset is primarily applied to address the inventor name disambiguation issue in patent data, thereby improving the quality and usability of patent data.