arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1102.3176cs.ITcs.LGmath.ITstat.ML
收藏

用最大逼近容量选择截断SVD的秩

Selecting the rank of truncated SVD by Maximum Approximation Capacity

ETH Zurich(苏黎世联邦理工学院)

Mario Frank, Joachim M. Buhmann

摘要:

截断奇异值分解(SVD)计算给定输入矩阵的最接近秩-$k$近似。在SVD的大多数应用中,选择适当的秩$k$定义了关键的模型阶数选择。为了获得一个原则性的截止准则的频谱,我们转换成一个嘈杂的信道编码问题的基本优化问题。该通道的最佳逼近能力控制正则化的适当强度以抑制噪声。在模拟实验中,这种信息理论方法来确定最佳的排名竞争与国家的最先进的模型选择技术。

英文摘要:

Truncated Singular Value Decomposition (SVD) calculates the closest rank-$k$ approximation of a given input matrix. Selecting the appropriate rank $k$ defines a critical model order choice in most applications of SVD. To obtain a principled cut-off criterion for the spectrum, we convert the underlying optimization problem into a noisy channel coding problem. The optimal approximation capacity of this channel controls the appropriate strength of regularization to suppress noise. In simulation experiments, this information theoretic method to determine the optimal rank competes with state-of-the art model selection techniques.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑