一种融入用户情绪因素的综合音乐推荐方法
发布时间：2018-12-29 14:03
【摘要】：随着互联网和移动互联网的发展,越来越多的用户在消费数字音像制品时,希望从繁多的内容中得到更好的推荐。不同于普通商品,音乐的消费过程与用户的情绪关联十分密切。本文设计了一种融入用户情绪因素的综合音乐推荐方法。该方法利用现代心理学中人类情绪的自然节律模型,结合时间变化来推测用户当前的主导情绪。通过人工情绪模型来增加情绪的维度,以产生粒度更精细的情绪,并通过用户输入的训练,来逼近用户当前的真实情绪。在现有的音乐推荐方法基础上,加入或过滤出情绪分类与用户情绪吻合的推荐结果。实验结果验证了情绪因素会对音乐推荐过程产生影响,构建的融入用户情绪因素的综合音乐推荐方法,比其他不考虑情绪因素的推荐策略好且稳定,具有较好的实际应用效果。
[Abstract]:With the development of Internet and mobile Internet, more and more users hope to get better recommendation from various content when consuming digital audio and video products. Unlike ordinary goods, the consumption process of music is closely related to the user's emotions. This paper designs a comprehensive music recommendation method which integrates user emotion factors. This method uses the natural rhythm model of human emotion in modern psychology and combines the time change to speculate the current dominant emotion of the user. The artificial emotion model is used to increase the dimension of emotion to produce more granular emotion and to approach the user's current real emotion through user input training. On the basis of the existing music recommendation methods, we add or filter out the recommended results of emotion classification and user emotion matching. The experimental results show that the emotional factors will affect the music recommendation process. The integrated music recommendation method which integrates the user emotional factors is better and more stable than other recommendation strategies without taking into account the emotional factors and has a better practical application effect.
【作者单位】： 浙江工商大学管理工程与电子商务学院;
【基金】：国家自然科学基金(71571162) 国家科技支撑计划项目(2014BAH24F06) 教育部人文社会科学重点研究基地项目(14JJD630011)
【分类号】：TP391.3
，
本文编号：2394922
[Abstract]:With the development of Internet and mobile Internet, more and more users hope to get better recommendation from various content when consuming digital audio and video products. Unlike ordinary goods, the consumption process of music is closely related to the user's emotions. This paper designs a comprehensive music recommendation method which integrates user emotion factors. This method uses the natural rhythm model of human emotion in modern psychology and combines the time change to speculate the current dominant emotion of the user. The artificial emotion model is used to increase the dimension of emotion to produce more granular emotion and to approach the user's current real emotion through user input training. On the basis of the existing music recommendation methods, we add or filter out the recommended results of emotion classification and user emotion matching. The experimental results show that the emotional factors will affect the music recommendation process. The integrated music recommendation method which integrates the user emotional factors is better and more stable than other recommendation strategies without taking into account the emotional factors and has a better practical application effect.
【作者单位】： 浙江工商大学管理工程与电子商务学院;
【基金】：国家自然科学基金(71571162) 国家科技支撑计划项目(2014BAH24F06) 教育部人文社会科学重点研究基地项目(14JJD630011)
【分类号】：TP391.3
，
本文编号：2394922
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2394922.html
