COOKIE
收藏资源简介:
COOKIE数据集是由罗格斯大学和HPI/波茨坦大学的研究团队开发的,专门用于电子商务平台中的对话推荐系统。该数据集通过整合用户与代理之间的对话以及定制的知识图谱来构建,旨在模拟用户从粗到细选择商品的过程。数据集包含超过200万条对话,覆盖多个电子商务领域,如手机配件、食品杂货、玩具和汽车等。创建过程中,研究团队首先构建了一个统一的知识图谱,然后提取关键实体以形成对话骨架,最后通过模拟对话过程来合成数据集。COOKIE数据集的应用领域主要集中在提升对话推荐系统的性能,通过丰富的知识图谱支持,帮助用户更有效地找到最适合的商品。
The COOKIE Dataset was developed by research teams from Rutgers University and HPI/University of Potsdam, specifically tailored for conversational recommendation systems in e-commerce platforms. This dataset is constructed by integrating conversations between users and agents along with a customized knowledge graph, aiming to simulate the process of users selecting products from coarse-grained to fine-grained levels. The dataset contains over 2 million conversations, covering multiple e-commerce sectors such as mobile phone accessories, groceries, toys, automobiles and more. During its creation, the research team first built a unified knowledge graph, then extracted key entities to form the conversation skeleton, and finally synthesized the dataset by simulating the conversation process. The main application scenarios of the COOKIE Dataset focus on improving the performance of conversational recommendation systems, leveraging the rich knowledge graph support to help users more effectively find the most suitable products.




