MirrorUser/UserMirrorer

UserMirrorrer-eval数据集是UserMirrorer的训练集，用于构建与用户偏好对齐的用户模拟器。数据集包含用户ID、项目ID、时间戳、选择的消息列表、拒绝的消息列表、印象列表和项目位置等特征。原始数据来自MIND和MovieLens-1M数据集，但由于许可限制，不能直接包含在内。README提供了详细的流程来加载原始数据、执行预处理操作并组装最终的训练和测试集。

UserMirrorrer-eval is the training set of UserMirrorer, designed for building preference-aligned user simulators. The dataset includes features such as user ID, item ID, timestamp, list of chosen messages, list of rejected messages, impression list, and item position. The raw data originates from the MIND and MovieLens-1M datasets but cannot be included directly due to licensing restrictions. The README provides a comprehensive pipeline to load the original data, perform necessary preprocessing, and assemble the final training and test splits.