研究发现年龄影响听歌偏好，年轻人爱流行音乐，老年人更个性化。
Soundtracks of Our Lives: How Age Influences Musical Preferences
- 用长期用户数据追踪听歌习惯随年龄变化
- 年轻用户偏好广泛听流行音乐，年长者更个性化
- 为长期推荐系统设计提供新方向，适合做用户建模的学者
推荐系统研究通常基于短期用户交互数据，但实际应用中系统会长期运行，用户行为也会随时间演变。尽管心理学和媒体研究已揭示年龄对偏好的影响，但在用户建模与推荐系统领域仍缺乏相关研究。本研究基于LFM-2b数据集——目前唯一具备足够长时跨度和用户年龄信息的数据集——分析了用户偏好与行为的演化规律。结果表明：年轻用户倾向于广泛收听当代流行音乐，而年长用户则表现出更复杂、个性化的听歌习惯。这些发现为推荐系统研究提供了重要启示，指明了未来发展方向。
原文摘要 · Abstract (English)
The majority of research in recommender systems, be it algorithmic improvements, context-awareness, explainability, or other areas, evaluates these systems on datasets that capture user interaction over a relatively limited time span. However, recommender systems can very well be used continuously for extended time. Similarly so, user behavior may evolve over that extended time. Although media studies and psychology offer a wealth of research on the evolution of user preferences and behavior as individuals age, there has been scant research in this regard within the realm of user modeling and recommender systems. In this study, we investigate the evolution of user preferences and behavior using the LFM-2b dataset, which, to our knowledge, is the only dataset that encompasses a sufficiently extensive time frame to permit real longitudinal studies and includes age information about its users. We identify specific usage and taste preferences directly related to the age of the user, i.e., while younger users tend to listen broadly to contemporary popular music, older users have more elaborate and personalized listening habits. The findings yield important insights that open new directions for research in recommender systems, providing guidance for future efforts.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。