改进的基于狄利克雷混合模型的推荐算法

doi:10.13229/j.cnki.jdxbgxb20161325

吉林大学学报(工学版) ›› 2018, Vol. 48 ›› Issue (2): 596-604.doi: 10.13229/j.cnki.jdxbgxb20161325

改进的基于狄利克雷混合模型的推荐算法

董坚峰^{1, 2}, 张玉峰³, 戴志强¹

1.吉首大学软件学院,湖南张家界427000;
2.中山大学管理学院, 广州510275;
3 武汉大学信息资源研究中心,武汉430072

收稿日期:2016-12-06 出版日期:2018-03-01 发布日期:2018-03-01
作者简介:董坚峰(1977-),男,副教授,在站博士后. 研究方向:计算机信息系统工程,网络信息资源管理.E-mail:yuanrele@163.com
基金资助:
国家自然科学基金项目(71373197); 湖南省哲学社会科学基金项目(14YBA318); 湖南省教育厅优秀青年项目(17B221)

Improved recommendation algorithm based on DPM model

DONG Jian-feng^{1, 2}, ZHANG Yu-feng³, DAI Zhi-qiang¹

1.School of Software,Jishou University,Zhangjiajie 427000,China;
2.Sun Yat-Sen Business School,Sun Yat-sen University,Guangzhou 510275,China;
3.Center for Studies of Information Resources,Wuhan University,Wuhan 430072,China

Received:2016-12-06 Online:2018-03-01 Published:2018-03-01

摘要/Abstract

摘要： 数据动态性在设计推荐算法过程中不能忽略。针对大多数传统型静态文本建模方法主要基于可交换性的基本假设,对数据在协变量空间上的依赖关系有所忽略的问题,本文提出了一种新的主要基于函数式DPM模型过程动态推荐模型。该模型对传统DPM混合模型在动态数据建模方面的问题进行了改进。创建了相关狄利克雷过程的参数与协变量空间联系,且狄利克雷过程仍然属于边际分布。应用函数式狄利克雷过程,可针对产生、消失以及参数改变的混合模型组件进行有效建模,并可作为动态先验融入非参数混合模型。仿真实验结果表明,与应用传统狄利克雷过程做先验的话题模型相比,本文算法优势更加明显。

关键词: 计算机系统结构, DPM模型, 动态数据, 吉布斯采样, 折叠采样推演算法

Abstract: Data dynamics can not be ignored in the design of recommendation algorithm. Most of the traditional static text modeling methods are based on the basic assumption of exchangeability, but the dependence of the data in the covariate space is neglected. To solve this problem, a new dynamic function DPM recommendation model is proposed based on process model. This model improves the traditional Diky mixture model in dynamic data modeling. The space relationship of the parameters and covariates of the dependent Dirichlet process is created, while the Dirichlet process still belongs to the marginal distribution. Application of functional Dirichlet process can carry out effective modeling for mixed model assembly for disappeared and the change of the parameters, and can be used as a dynamic prior into parametric mixture model. Simulation results show that, compared with traditional de Lickley process, the proposed algorithm has more obvious advantages in doing a priori topic model.

Key words: computer systems organization, DPM model, dynamic data, Gibbs sampling, folding sampling deduction algorithm

中图分类号:

TP393

董坚峰, 张玉峰, 戴志强. 改进的基于狄利克雷混合模型的推荐算法[J]. 吉林大学学报(工学版), 2018, 48(2): 596-604.

DONG Jian-feng, ZHANG Yu-feng, DAI Zhi-qiang. Improved recommendation algorithm based on DPM model[J]. 吉林大学学报(工学版), 2018, 48(2): 596-604.

参考文献

[1] 解男男, 胡亮, 努尔布力,等. 基于Web日志挖掘的网页推荐方法[J]. 吉林大学学报:理学版, 2013, 51(2):267-272.
Xie Nan-nan,Hu Liang,Nurbolz,et al. Web recommender system based on Web lob minim[J].Journal of Jilin University(Science Edition), 2013,51(2):267-272.
[2] 董立岩, 王越群, 贺嘉楠,等. 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.
Dong Li-yan, Wang Yue-qun, He Jia-nan, et al.Collaborative filtering recommendation algorithm based on time decay[J].Journal of Jilin University(Engineering and Technology Edition) 2017, 47(4):1268-1272.
[3] Jiang J, Lu J, Zhang G, et al. Scaling-up item-based collaborative filtering recommendation algorithm based on hadoop[C]∥ Services, IEEE, 2011:490-497.
[4] Chen W, Niu Z, Zhao X, et al. A hybrid recommendation algorithm adapted in e-learning environments[J]. World Wide Web, 2014, 17(2):271-284.
[5] Qiu T, Chen G, Zhang Z K, et al. An item-oriented recommendation algorithm on cold-start problem[J]. EPL, 2011, 95(5):58003.
[6] Lin K, Wang J, Wang M, et al. A hybrid recommendation algorithm based on Hadoop[D]. Institute of Electrical and Electronics Engineers Inc. 2014.
[7] Zhang J, Peng Q, Sun S, et al. Collaborative filtering recommendation algorithm based on user preference derived from item domain features[J]. Physica A Statistical Mechanics & Its Applications, 2014, 396(2):66-76.
[8] Zhong Z, Sun Y, Wang Y, et al. An improved collaborative filtering recommendation algorithm not based on item rating[C]∥IEEE, International Conference on Cognitive Informatics & Cognitive Computing. IEEE, 2015:230-233.
[9] Lu Z, Shen H. A security-assured accuracy-maximised privacy preservingcollaborative filtering recommendation algorithm[C]∥Proceedings of the 19th International Database Engineering & Applications Symposium, Yokohama, Japan,2015:72-80.
[10] Huang Y M, Kuo Y H, Chen J N, et al. NP-miner: a real-time recommendation algorithm by using web usage mining[J]. Knowledge-Based Systems, 2006, 19(4):272-286.
[11] 邓爱林, 左子叶, 朱扬勇. 基于项目聚类的协同过滤推荐算法[J]. 小型微型计算机系统, 2004, 25(9):1665-1670.
Deng Ai-lin, Zuo Zi-ye, Zhu Yang-yong. Collaborative filtering recommendation algorithm based on item clustering [J].Mimi-micro Systems,2004,25(9):1665-1670.
[12] Huang Z, Zeng D, Chen H. A comparison of collaborative-filtering recommendation algorithms for E-commerce[J]. Intelligent Systems of IEEE, 2007, 22(5):68-78.
[13] 严宇宇, 陶煜波, 林海. 基于层次狄利克雷过程的交互式主题建模[J]. 软件学报, 2016(5):1114-1126.
Yan Yu-yu,Tao Yu-bo, Lin Hai.Interactive topic modeling based on hierarchical Dirichlet Process [J].Journal of Software,2016(5):1114-1126.
[14] Caron F, Davy M, Doucet A. Generalized polya urn for time-varying Dirichlet process mixtures[C]∥Proceedings of Proc of UAI’07, Corvallis, Oregon, 2007:33-40.
[15] Caron F, Davy M, Doucet A. Generalized Polya Urn for time-varying Dirichlet process mixtures[J].2012.
[16] Rabaoui A, Viandier N, Duflos E, et al. DPMs for the density estimation in a dynamic nonlinear modeling: application to GPS positioning in urban canyons[J]. IEEE Transactions on Signal Processing, 2012, 60:1638-1655.
[17] Blei D M, Frazier P. Distance dependent Chinese restaurant processes[C]∥Proceedings of Proc of ICML’10, Haifa, Israel, 2010:87-94.
[18] Zhang M, Hurley N. Avoiding monotony: improving the diversity of recommendation lists[C]∥Proceedings of Proceedings of the 2008 ACM Conference on Recommender Systems,ACM,2008:123-130.
[19] Xu M, Zhu J, Zhang B. Fast max-margin matrix factorization with data augmentation[C]∥Proceedings of the 30th International Conference on Machine Learning (ICML-13),2013:978-986.
[20] Li L, Chu W, Langford J, et al. Unbiased offline evaluation of contextual-bandit-based news article recommendation algorithms[J]. Wsdm, 2012:297-306.

相关文章 15

[1]	余宜诚, 胡亮, 迟令, 初剑峰. 一种改进的适用于多服务器架构的匿名认证协议[J]. 吉林大学学报(工学版), 2018, 48(5): 1586-1592.
[2]	赵博, 秦贵和, 赵永哲, 杨文迪. 基于半陷门单向函数的公钥密码[J]. 吉林大学学报(工学版), 2018, 48(1): 259-267.
[3]	刘磊, 刘利娟, 吴新维, 张鹏. 基于ECPMR的编译器测试方法[J]. 吉林大学学报(工学版), 2017, 47(4): 1262-1267.
[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]	于斌斌, 武欣雨, 初剑峰, 胡亮. 基于群密钥协商的无线传感器网络签名协议[J]. 吉林大学学报(工学版), 2017, 47(3): 924-929.
[6]	邓昌义, 郭锐锋, 张忆文, 王鸿亮. 基于平衡因子的动态偶发任务低功耗调度算法[J]. 吉林大学学报(工学版), 2017, 47(2): 591-600.
[7]	魏晓辉, 刘智亮, 庄园, 李洪亮, 李翔. 支持大规模流数据在线处理的自适应检查点机制[J]. 吉林大学学报(工学版), 2017, 47(1): 199-207.
[8]	郝娉婷, 胡亮, 姜婧妍, 车喜龙. 基于多管理节点的乐观锁协议[J]. 吉林大学学报(工学版), 2017, 47(1): 227-234.
[9]	魏晓辉, 李翔, 李洪亮, 李聪, 庄园, 于洪梅. 支持大规模流数据处理的弹性在线MapReduce模型及拓扑协议[J]. 吉林大学学报(工学版), 2016, 46(4): 1222-1231.
[10]	车翔玖, 梁森. 一种基于大顶堆的SPIHT改进算法[J]. 吉林大学学报(工学版), 2016, 46(3): 865-869.
[11]	董悦丽, 郭权, 孙斌, 康玲. 药物分子对接动态任务迁移优化[J]. 吉林大学学报(工学版), 2015, 45(4): 1253-1259.
[12]	匡哲君，师唯佳，胡亮. 基于无线传感器网络的角色成员关系剩余能量新算法[J]. 吉林大学学报(工学版), 2015, 45(2): 600-605.
[13]	张忆文，郭锐锋. 实时系统混合任务低功耗调度算法[J]. 吉林大学学报(工学版), 2015, 45(1): 261-266.
[14]	张忆文^{1, 2}, 郭锐锋¹. 制的容错节能调度算法[J]. 吉林大学学报(工学版), 2014, 44(4): 1112-1117.
[15]	付帅¹, 马建峰¹, 李洪涛¹, 王长广². 改进的基于分簇无线传感器网络的数据聚合算法[J]. 吉林大学学报(工学版), 2014, 44(4): 1118-1125.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

改进的基于狄利克雷混合模型的推荐算法

Improved recommendation algorithm based on DPM model

RICH HTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 0