REASONER
收藏资源简介:
REASONER数据集由北京大数据管理和分析方法重点实验室创建,旨在通过多方面的真实用户标注来评估推荐系统的解释性。数据集包含约3000名用户的反馈,这些用户在视频推荐平台上回答了一系列关于推荐解释性的问题。数据集不仅涵盖了文本解释,还包括视觉解释,适用于多种推荐解释任务。创建过程中,通过精心设计的问题和规则来确保数据质量,旨在解决推荐系统解释性的量化评估问题,为推荐系统领域提供新的研究机会。
The REASONER dataset was developed by the Beijing Key Laboratory of Big Data Management and Analysis Methods, aiming to evaluate the explainability of recommendation systems via multi-faceted real user annotations. The dataset contains feedback from approximately 3,000 users who answered a series of questions regarding recommendation explainability on a video recommendation platform. It covers not only textual explanations but also visual explanations, making it applicable to various recommendation explanation tasks. During its creation, carefully designed questions and rules were adopted to ensure data quality, addressing the challenge of quantitative evaluation for recommendation system explainability and providing new research opportunities for the field of recommendation systems.

- 1REASONER: An Explainable Recommendation Dataset with Multi-aspect Real User Labeled Ground Truths Towards more Measurable Explainable Recommendation北京大数据管理和分析方法重点实验室 · 2023年



