基于信息检索及k均值聚类的音乐个性化推荐算法

doi:10.13229/j.cnki.jdxbgxb20200243

摘要/Abstract

摘要：

为追求更高的推荐质量与更快的推荐效率，以音乐资源为研究目标，构建了一种结合MIR与k-means标签聚类的个性化推荐算法。选取音高变化与轮廓作为系统处理数据，利用输入、预处理、特征提取、相似度匹配和输出模块架构音乐信息检索系统，依据用户与资源的选择关系形成多模网络，通过延伸、局域范围界定以及连接优先化阶段组建检索系统的动态多维网络模型，采用k-means标签聚类搜索邻域用户，获取最近邻用户集，将解得的综合特征值设定成初始聚类中心，根据排序后的推荐资源预测结果，实现个性化推荐。针对阿里天池音乐数据开展仿真实验，采用平均绝对误差、准确率和召回率等指标验证算法性能，经实验数据比对后发现，本文算法准确率较为理想，误差较小，具备比较优越的推荐有效性。

关键词: 音乐信息检索, k-means算法, 标签聚类, 音高轮廓, 个性化推荐, 相似度

Abstract:

In order to pursue higher recommendation quality and faster recommendation efficiency， a personalized recommendation algorithm based on Mir and k-means tag clustering is constructed with music resources as the research objective. The pitch change and contour are selected as the system processing data， and the input， preprocessing， feature extraction， similarity matching and output modules are used to construct the music information retrieval system. According to the selection relationship between users and resources， a multi-mode network is formed. The dynamic multi-dimensional network model of the retrieval system is established through the extension， local range definition and connection priority stages Tag clustering searches neighborhood users to obtain the nearest neighbor user set. The comprehensive eigenvalue is set as the initial clustering center， and personalized recommendation is realized according to the sorted recommendation resource prediction results. The performance of the algorithm is verified by the average absolute error， accuracy rate and recall rate. After the comparison of experimental data， it is found that the accuracy of the proposed algorithm is ideal， the error is small， and the recommendation effectiveness is superior.

Key words: music information retrieval, k-means algorithm, label clustering, pitch contour, personalized recommendation, similarity

中图分类号:

TP393

王海龙,柳林,林民,裴冬梅. 基于信息检索及k均值聚类的音乐个性化推荐算法[J]. 吉林大学学报(工学版), 2021, 51(5): 1845-1850.

Hai-long WANG,Lin LIU,Min LIN,Dong-mei PEI. Music personalized recommendation algorithm based on k⁃means clustering algorithm[J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(5): 1845-1850.

图/表 5

图1

表1

表2

表3

图2

参考文献 20

1	Minerva B B, Dolores G R, Isidora F F, et al. Use of new technologies by adolescents in the search for health information[J]. Atencion Primaria, 2018, 50(9): 547-552.
2	Jia M, Komeily A, Wang Y, et al. Adopting internet of things for the development of smart buildings: a review of enabling technologies and applications[J]. Automation in Construction, 2019, 101(5): 111-126.
3	Im H, Song H, Jung J. The effect of streaming services on the concentration of digital music consumption[J]. Information Technology & People, 2019, 33(1): 160-179.
4	马闻锴, 李贵, 李征宇,等. 一种基于标签的Top-N个性化推荐算法[J]. 计算机科学, 2019, 46(): 224-229.
	Ma Wen-kai, Li Gui, Li Zheng-yu, et al. A Top-N personalized recommendation algorithm based on tags[J]. Computer Science, 2019, 46(Sup.2): 224-229.
5	张鹏飞, 王宜贵, 张志军. 融合标签和多元信息的个性化推荐算法研究[J]. 计算机工程与应用, 2019, 55(5): 159-165.
	Zhang Peng-fei, Wang Yi-gui, Zhang Zhi-jun. Research on personalized recommendation algorithm combining tags and multiple information[J]. Computer Engineering and Applications, 2019, 55(5): 159-165.
6	Davis R F, Baese-Berk M M, Stilp C. Effects of pitch contour and speaking rate on perception of foreign-accented speech[J]. The Journal of the Acoustical Society of America, 2019, 145(3): 1827.
7	Dewan S, Ho Y J, Ramaprasad J. Popularity or proximity: characterizing the nature of social influence in an online music community[J]. Operations Research, 2018, 58(3): 267-270.
8	Jain B J, Schultz D. Asymmetric learning vector quantization for efficient nearest neighbor classification in dynamic time warping spaces[J]. Pattern Recognition, 2018, 76: 349-366.
9	李宾, 周旭, 梅芳, 等. 基于K-means和矩阵分解的位置推荐算法[J]. 吉林大学学报: 工学版, 2019, 49(5): 1653-1660.
	Li Bin, Zhou Xu, Mei Fang, et al. Location recommendation algorithm based on K-means and matrix factorization[J]. Journal of Jilin University(Engineering and Technology Edition), 2019, 49(5):1653-1660.
10	Wang X. Research on the improved method of fundamental frequency extraction for music automatic recognition of piano music[J]. Journal of Intelligent and Fuzzy Systems, 2018, 35(3): 1-7.
11	Dazhi X U. Research on music culture personalized recommendation based on factor decomposition machine[J]. Personal and Ubiquitous Computing, 2020, 24(2): 247-257.
12	刘洲洲, 尹文晓, 张倩昀, 等. 基于离散优化算法和机器学习的传感云入侵检测[J]. 吉林大学学报: 工学版, 2020, 50(2): 692-702.
	Liu Zhou-zhou, Yin Wen-xiao, Zhang Qian-yun, et al. Sensor cloud intrusion detection based on discrete optimization algorithm and machine learning[J]. Journal of Jilin University(Engineering and Technology Edition), 2020, 50(2): 692-702.
13	张璠, 李为相, 李为. 基于从众心理矩阵和加权相似度的推荐算法[J]. 计算机工程与设计, 2019, 40(6): 1616-1620.
	Zhang Pan, Li Wei-xiang, Li Wei. Recommendation algorithm based on herd mental matrix and weighted similarity[J]. Computer Engineering and Design, 2019, 40(6): 1616-1620.
14	王瑶, 寇月, 申德荣, 等. 元路径选择和矩阵分解的跨社交网络链路预测[J]. 计算机科学与探索, 2019, 13(9): 1459-1470.
	Wang Yao, Kou Yue, Shen De-rong, et al. Cross-social network link prediction based on meta-path selection and matrix decomposition[J]. Journal of Computer Science and Exploration, 2019, 13(9): 1459-1470.
15	周本金, 陶以政, 纪斌, 等. 最小化误差平方和k-means初始聚类中心优化方法[J]. 计算机工程与应用, 2018, 54(15): 48-52.
	Zhou Ben-jin, Tao Yi-zheng, Ji Bin, et al. K-means initial clustering center optimization method to minimize the error square sum[J]. Computer Engineering and Applications, 2018, 54(15): 48-52.
16	张铁赢, 黄贵, 章颖强, 等. X-DB:软硬一体的新型数据库系统[J]. 计算机研究与发展, 2018, 55(2): 319-326.
	Zhang Tie-ying, Huang Gui, Zhang Ying-qiang, et al. X-DB: a new type of database system integrating software and hardware[J]. Journal of Computer Research and Development, 2018, 55(2): 319-326.
17	徐毅, 叶卫根, 戴鑫, 等. 融合用户信任度与相似度的推荐算法研究[J]. 小型微型计算机系统, 2018, 39(1): 78-83.
	Xu Yi, Ye Wei-gen, Dai Xin, et al. Research on recommendation algorithm combining user trust and similarity[J]. Small Microcomputer System, 2018, 39(1): 78-83.
18	Xia J, Li G, Cao Z, et al. Personalized exercise recommendation algorithm combining learning objective and assignment feedback[J]. Journal of Intelligent & Fuzzy Systems, 2018, 35(3): 2965-2973.
19	Cao B, Chen J, Liu J, et al. A topic attention mechanism and factorization machines based mobile application recommendation method[J]. Mobile Networks and Applications, 2020(4): 1-12.
20	赖奕安, 张玉洁, 杜雨露, 等. 一种基于协同上下文关系学习的同城活动推荐算法[J]. 软件学报, 2020, 31(2): 421-438.
	Lai Yi-an, Zhang Yu-jie, Du Yu-lu, et al. A recommendation algorithm for intra-city activities based on collaborative context learning[J]. Journal of Software, 2020, 31(2): 421-438.

相关文章 15

[1]	桂春, 黄旺星. 基于改进的标签传播算法的网络聚类方法[J]. 吉林大学学报(工学版), 2018, 48(5): 1600-1605.
[2]	王旭, 欧阳继红, 陈桂芬. 基于垂直维序列动态时间规整方法的图相似度度量[J]. 吉林大学学报(工学版), 2018, 48(4): 1199-1205.
[3]	王旭, 欧阳继红, 陈桂芬. 基于多重序列所有公共子序列的启发式算法度量多图的相似度[J]. 吉林大学学报(工学版), 2018, 48(2): 526-532.
[4]	董立岩, 王越群, 贺嘉楠, 孙铭会, 李永丽. 基于时间衰减的协同过滤推荐算法[J]. 吉林大学学报(工学版), 2017, 47(4): 1268-1272.
[5]	王贵参, 黄岚, 王岩, 宋立明, 欧歌. 引入极值非相邻连接的连接聚类方法[J]. 吉林大学学报(工学版), 2016, 46(5): 1616-1621.
[6]	王克朝, 王甜甜, 苏小红, 马培军. 基于频繁闭合序列模式挖掘的学生程序雷同检测[J]. 吉林大学学报(工学版), 2015, 45(4): 1260-1265.
[7]	王玉磊, 赵春晖, 齐滨. 基于光谱相似度量的高光谱图像异常检测算法[J]. 吉林大学学报(工学版), 2013, 43(增刊1): 148-153.
[8]	苑文举, 王健. E-business信任关系网建模[J]. 吉林大学学报(工学版), 2011, 41(增刊2): 259-263.
[9]	吕丹,毕笃彦. 基于结构相似的DCT域图像质量评价[J]. 吉林大学学报(工学版), 2011, 41(6): 1771-1776.
[10]	郭耸, 顾国昌, 蔡则苏, 刘海波, 沈晶. 基于肤色分割和改进的AdaBoostSVM算法的人脸检测[J]. 吉林大学学报(工学版), 2011, 41(02): 473-0478.
[11]	田野, 刘大有. 改进的PeertoPeer环境下的聚类算法[J]. 吉林大学学报(工学版), 2010, 40(06): 1639-1643.
[12]	刘彦斌, 周春光, 张重毅, 周栩, 翟延冬, 邹淑雪. 基于本体的语义生物信息数据库资源链接[J]. 吉林大学学报(工学版), 2010, 40(06): 1650-1654.
[13]	刘志勇,刘磊,刘萍萍,杨帆,贾冰. 一种基于语义网的个性化学习资源推荐算法[J]. 吉林大学学报(工学版), 2009, 39(增刊2): 391-0395.
[14]	刘春辰，，刘大有，，王生生，，赵静滨，，王兆丹， . 改进的语义相似度计算模型及应用[J]. 吉林大学学报(工学版), 2009, 39(01): 119-123.
[15]	杨印生;孙赵华;马萍;陶跃;司瑾 . 基于SOMK算法的T-S模糊系统建模方法[J]. 吉林大学学报(工学版), 2008, 38(03): 658-0661.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

喜爱程度	评分	判定标准
最喜欢	5	标记为红心
不讨厌	3	听完未做标记
不喜欢	2	播放下一首
讨厌	1	标记为不再推荐

指标	100 kbit	1 Mbit	10 Mbit
评分数量/条	95 428	1 036 573	11 873 482
用户数量/人	9 357	5 823	713 578
音乐数量/首	1 882	4 327	112 837

文件名称	数据种类	格式
rating.dat	评分数据	User ID：：Music ID：：Rating：： Time stamp
music.dat	音乐数据	Music ID：：Title：：Genres
user.dat	用户数据	User ID：：age：：gender：：occupation：：zip code