摘要
为了丰富网络音乐检索系统的检索方式,提高检索结果的准确率,提出了一种基于深度学习的网络音乐检索系统,引入伪相关反馈技术扩展查询内容,完成查询结果的初排序,在此基础上结合用户提供的社会信息再次对查询结果进行合理的排序,最终通过排序学习获得质量最高的查询结果。实际应用结果表明,所设计的系统查询结果的准确性高于其他网络音乐检索系统,具有很强的实用性。
In order to enrich the retrieval methods of network music retrieval system and improve the accuracy of retrieval results,a network music retrieval system based on deep learning is proposed.Pseudo relevance feedback technology is introduced to expand the query content and complete the initial sorting of the query results.On this basis,the query results are sorted reasonably according to the social information provided by users.Order learning is used to obtain the highest quality query results.The practical application results show that the accuracy of the query results of the system designed in this paper is higher than that of other network music retrieval systems,and has strong practicability.
作者
黄文专
HUANG Wenzhuan(College of Stomatology Zhaoqing Medical College, Zhaoqing 526020, China)
出处
《微型电脑应用》
2022年第6期177-179,共3页
Microcomputer Applications
关键词
伪相关反馈技术
检索系统
深度学习
网络音乐
pseudo relevance feedback technology
retrieval system
deep learning
network music
