鲁棒主成分分析算法综述
首发时间:2016-03-09
摘要:主成分分析(principle component analysis)是对高维数据进行处理、分析、压缩以及可视化的一个流行工具。在网页查询、计算机视觉中的生物信息应用、图像分析等方面有广泛的应用。但是在现实场景中的应用和表现往往会受外点和受损的观察数据等的影响,使其表现不尽如人意。因此增强主成分分析的鲁棒性就显得尤为重要.John Wright等人提出的鲁棒主成分分析模型是目前最流行的模型.本文针对John Wright等人提出的鲁棒主成分分析模型,总结了近年来比较实用的几个算法。通过模拟实验对这些算法的运行效果和效率进行了对比。并在最后给出了鲁棒主成分分析在背景分离方面的一个应用。
关键词: 鲁棒主成分分析;迭代阈值算法;加速近端梯度法;增广拉格朗日算法
For information in English, please click here
survey on algorithms of robust principal component analysis
Abstract:Principal Component Analysis (PCA) is a popular tool for high-dimensional data processing, analysis,compression,and visualization. And it has wide applications in web search, bioinformatics, image analysis and so on. However, this model usually breaks down in the real applications because of the outliers and missing data. So it is important to enhance the robust of the PCA model. To achieve this, John Wright etc. proposed the robust principal component analysis model, which is the most popular model at present. This paper makes a brief survey on the existing algorithms of robust principal component analysis and compares the results and running speed with each other by numerical tests. At last, we also give an application about RPCA in background subtraction.
Keywords: Robust Principal Component Analysis the Iterative Thresholding Approach the Accelerated Proximal Gradient Approach the Methods of Augmented Lagrange Multipliers
基金:
论文图表:
引用
No.4679579113631314****
同行评议
共计0人参与
勘误表
鲁棒主成分分析算法综述
评论
全部评论0/1000