Dataset of predictions of which US patents are related to Artificial Intelligence

This is the dataset of predictions for which US patents belong to AI, as part of the research paper entitled "Patent Landscaping Using Document-Class Similarity: The Case of Artificial Intelligence" by Reza Rezazadegan and Mehran Moein. For each patent an "AI relevance score" is computed and reported, as well as the prediction of whether it is an AI patent or not. For the ones which belong to AI, a prediction of which AI subdomain the patent belongs to is made as well.本数据集为Reza Rezazadegan与Mehran Moein发表的题为《基于文档类别相似度的专利景观分析：以人工智能（Artificial Intelligence）领域为例》的研究论文的配套数据集，用于预测美国专利的人工智能归属。针对每一件美国专利，本数据集均计算并报告其人工智能关联度得分，同时给出该专利是否属于人工智能专利的预测结果。对于被判定为人工智能专利的样本，本数据集还进一步预测了该专利所属的人工智能子领域。