###
Journal of Software:2015.26(9):2262-2277

面向时序数据的矩阵分解
黄晓宇,潘嵘,李磊,梁冰,陈康,蔡文学
(华南理工大学 经济与贸易学院, 广东 广州 510006;中山大学 计算机软件研究所, 广东 广州 510275;中国电信股份有限公司 广东研究院, 广东 广州 510630)
Matrix Factorization for Time Series Data
HUANG Xiao-Yu,PAN Rong,LI Lei,LIANG Bing,CHEN Kang,CAI Wen-Xue
(School of Economics and Commerce, South China University of Technology, Guangzhou 510006, China;Software Institute, Sun Yet-San University, Guangzhou 510275, China;Academy of Guangdong Telecom Company, China Telecom Corporation Limited, Guangzhou 510630, China)
Abstract
Chart / table
Reference
Similar Articles
Article :Browse 2034   Download 2636
Received:January 22, 2013    Revised:July 09, 2014
> 中文摘要: 研究一类特殊的矩阵分解问题:对由多个对象在一组连续时间点上产生的数据构成的矩阵R,寻求把它近似地分解为两个低秩矩阵UV的乘积,即RUT×V.有为数众多的时间序列分析问题都可归结为所研究问题的求解,如金融数据矩阵的因子分析、缺失交通流数据的估计等.提出了该问题的概率图模型,进而由此导出了其约束优化模型,最终给出了模型的求解算法.在不同的数据集上进行实验验证了该模型的有效性.
Abstract:The paper studies a matrix factorization problem for time series data, where the target matrix R consists of the equal length time series data generated by a set of objects. The goal is to find two low rank matrices U and V, such that RUT×V. Many time series analysis problems, such as finance data analysis and missing traffic data imputation, can be reduced to the proposed model. A probabilistic graphical representation for the problem is proposed, and a constrained optimization model from the graphical representation is derived. The solution algorithms for the proposed model is also presented. Empirical studies show that the proposed model is superior to the baselines.
文章编号:     中图分类号:    文献标志码:
基金项目:国家高技术研究发展计划(863)(2012AA12A203); 国家自然科学基金(61003140); 国家社会科学基金(13BTJ005) 国家高技术研究发展计划(863)(2012AA12A203); 国家自然科学基金(61003140); 国家社会科学基金(13BTJ005)
Foundation items:
Reference text:

黄晓宇,潘嵘,李磊,梁冰,陈康,蔡文学.面向时序数据的矩阵分解.软件学报,2015,26(9):2262-2277

HUANG Xiao-Yu,PAN Rong,LI Lei,LIANG Bing,CHEN Kang,CAI Wen-Xue.Matrix Factorization for Time Series Data.Journal of Software,2015,26(9):2262-2277