摘要
互联网技术和电子信息技术的迅速发展为整个时代提供了巨大的计算能力,个性化推荐系统成为时代产物的缩影。结合常用的推荐系统核心算法,设计了一种针对个性化音乐的Apriori改进算法,此算法通过用户信息进行深度学习,利用候选矩阵压缩的方法进行推荐优化,采用准确性、召回率等参数作为评价标准。以Last.fm音乐网站的部分数据作为分析样本,对选定音乐按个性化音乐推荐方式进行试验,Apriori改进算法在准确率和召回率方面均得到优化,推荐效果更优。在考虑推荐数量的前提下,Apriori改进算法的准确率和召回率均高于Plaucount算法,而相似度方面低于Plaucount算法。
Rapid development of the Internet technology and electronic information technology has provided huge computingpower for the whole era,and personalized recommendation system has become the epitome of the product of the era.Combined with the common core algorithm of recommendation system,this paper provides an improved Apriori algorithm for personalized music.This algorithm applies user information for in-depth learning,candidate matrix compression for recommendation optimization,accuracy,recall rate and other parameters as evaluation criteria.Takingpart of the data of Last.fm music Website as the analysis sample,the selected music is tested accordingto the personalized music recommendation mode.The Apriori improved algorithm is optimized in accuracy and recall rate,and the recommendation effect is better.On the premise of consideringthe number of recommendations,the accuracy and recall rate of Apriori improved algorithm are higher than that of Plaucount algorithm,and the similarity is lower than Plaucount algorithm.
作者
余莉娟
YU Lijuan(College of Art,Shangluo College,Shangluo 726000,China)
出处
《微型电脑应用》
2020年第10期140-143,共4页
Microcomputer Applications
关键词
深度学习
推荐系统
个性化
音乐
deep learning
recommendation system
personalization
music
