oyku-tugana/amazon-musical-instruments-2018-2023-5core
收藏资源简介:
该数据集是Amazon Reviews 2023数据集中音乐乐器类别的采样子集,专为数据挖掘课程推荐系统项目创建。采样策略包括:时间过滤(2018年1月1日至2023年9月12日)、迭代5核过滤(确保每个用户和物品至少有5次交互,直到稳定)、冷启动测试集(保留5000名用户,每人有1-2条评论用于冷启动评估)和随机种子42。数据集包含三个配置:interactions(交互数据)、items(物品数据)和cold_start(冷启动数据)。主要统计信息:25,746名用户、13,575个物品、222,276次交互、稀疏度0.999364、平均评分4.451,以及物品带有图像、描述和价格的比例。模式方面,交互/冷启动配置包含用户ID、物品ASIN、评分、时间戳等字段,物品配置包含物品ASIN、标题、描述、价格等字段。
This dataset is a sampled subset of the Musical Instruments category from the Amazon Reviews 2023 dataset, specially developed for recommender system projects in data mining courses. The sampling strategies include: temporal filtering (covering the period from January 1, 2018 to September 12, 2023), iterative 5-core filtering (ensuring each user and item has at least 5 interactions until the process stabilizes), a cold-start test set (retaining 5,000 users, each with 1-2 reviews for cold-start evaluation), and a fixed random seed of 42. The dataset includes three configurations: interactions (interaction data), items (item data), and cold_start (cold-start data). The main statistical metrics are: 25,746 users, 13,575 items, 222,276 interactions, a sparsity of 0.999364, an average rating of 4.451, and the proportion of items with images, descriptions and prices. For data schemas, the interactions and cold_start configurations contain fields such as user ID, item ASIN, rating, timestamp, etc., while the items configuration includes fields such as item ASIN, title, description, price, etc.




