公共环境下的混合型音乐推荐系统的关键技术研究

Hybrid music recommender system in public environment

随着计算机网络和多媒体技术的迅速普及,数字音乐消费已经成为人们日常生活中的常见活动,音乐推荐系统也因此成为了推荐系统和电子商务领域的一大研究热点。在对现有音乐推荐系统调研的基础上,重点研究公共环境下的混合型音乐推荐系统的设计和实现;将音乐特征和语境信息相结合,提出了一种新颖的混合型音乐推荐算法。为保证实际应用环境中音乐消费行为的灵活性,系统实现了投票和DJ两种推荐模式。该系统在实验室及实地测试中均取得了较高的用户满意度。

With the rapid development of the Internet and multimedia technologies,digital music consumptions have become daily activities.Music recommender systems（MRS） have thus become one of the key research issues in the fields of recommender systems and E-commerce.Based on a thorough investigation of existing MRSs,this research explored hybrid MRS in public environment.Based on a combination of musical features and contextual information,this paper presented a hybrid recommendation algorithm.And it presented voted and DJ modes to ensure flexibility in a practical scenario.Satisfactory results were achieved in the laboratory test and field study.