The emergence of Internet music has slowed down the restrictions of space and time on people’s enjoyment of music information services. However, in the face of massive and growing music works, information overload has become the most direct problem, and the need to improve user experience has become very urgent. One of the effective solutions to information overload is recommender system, which can help people to discover the interesting content from the complicated information. Therefore, the combination of recommendation system and Internet music has become an inevitable trend of music development. Referring to the traditional music recommendation methods, this paper proposes a big data music personalized recommendation method based on big data analysis, which combines user behavior, behavior context, user information, and music work information. In this paper, the user big data is introduced into the model building process. Through the factor decomposition machine (FM) learning method, the effect of various influencing factors on user behavior is analyzed to build the user dynamic interest model and complete the user preference acquisition. In the stage of recommendation candidate set selection, combining with the traditional collaborative filtering recommendation idea, the work of recommendation candidate set selection is carried out from two aspects. At the same time, this paper designed and completed a comparative experiment with the processing performance, accuracy and coverage as indicators, and verified the effectiveness of the improved recommendation method.

中文翻译：

---

基于因子分解机的音乐文化个性化推荐研究

互联网音乐的出现放慢了人们享受音乐信息服务的时空限制。然而，面对庞大且不断增长的音乐作品，信息过载已成为最直接的问题，并且改善用户体验的需求变得非常紧迫。推荐系统是解决信息过载的有效方法之一，它可以帮助人们从复杂的信息中发现有趣的内容。因此，推荐系统与网络音乐的结合已成为音乐发展的必然趋势。结合传统音乐推荐方法，提出了一种基于大数据分析的大数据音乐个性化推荐方法，该方法结合了用户行为，行为上下文，用户信息和音乐作品信息。本文将用户大数据引入到模型构建过程中。通过因子分解机学习方法，分析了各种影响因素对用户行为的影响，建立了用户动态兴趣模型，完成了用户偏好的获取。在推荐候选集选择阶段，结合传统的协同过滤推荐思想，从两个方面进行推荐候选集的选择工作。同时，设计并完成了以处理性能，准确性和覆盖率为指标的对比实验，验证了改进推荐方法的有效性。通过因子分解机（FM）学习方法，分析了各种影响因素对用户行为的影响，建立了用户动态兴趣模型，完成了用户偏好获取。在推荐候选集选择阶段，结合传统的协同过滤推荐思想，从两个方面进行推荐候选集的选择工作。同时，设计并完成了以处理性能，准确性和覆盖率为指标的对比实验，验证了改进推荐方法的有效性。通过因子分解机（FM）学习方法，分析了各种影响因素对用户行为的影响，建立了用户动态兴趣模型，完成了用户偏好获取。在推荐候选集选择阶段，结合传统的协同过滤推荐思想，从两个方面进行推荐候选集的选择工作。同时，设计并完成了以处理性能，准确性和覆盖率为指标的对比实验，验证了改进推荐方法的有效性。在推荐候选集选择阶段，结合传统的协同过滤推荐思想，从两个方面进行推荐候选集的选择工作。同时，设计并完成了以处理性能，准确性和覆盖率为指标的对比实验，验证了改进推荐方法的有效性。在推荐候选集选择阶段，结合传统的协同过滤推荐思想，从两个方面进行推荐候选集的选择工作。同时，设计并完成了以处理性能，准确性和覆盖率为指标的对比实验，验证了改进推荐方法的有效性。