matthewfranglen/lastfm-1k

这是一个包含用户、时间戳、艺术家和歌曲记录的音乐推荐系统训练数据集。用户可能提供了性别、年龄和位置等额外信息，但这些信息是可选的且未经验证。数据集已经去除了缺少musicbrainz标识符的行，并且提供了训练、验证和测试数据集，其中用户没有重叠。为了方便训练，数据集还包含了用户索引、艺术家索引和轨道索引列。

This dataset contains user, timestamp, artist, and song records for training a music recommendation system. Users may have provided additional information such as gender, age, and location, which are optional and unverified. The dataset has been cleaned to remove rows with missing musicbrainz identifiers, and it includes separate train, validation, and test sets without overlapping users. For ease of training, the dataset also includes user_index, artist_index, and track_index columns.