Can we use PCA to detect small signals in noisy data?

详细信息    查看全文

• 作者：Jakob Spiegelberg ; Já ; n Rusz
• 关键词：PCA ; Nullspace based denoising ; Blind source separation ; Bias estimation ; Spectrum imaging
• 刊名：Ultramicroscopy
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：172
• 期：Complete
• 页码：40-46
• 全文大小：875 K
• 卷排序：172

文摘

The influence of noise and size of the data set on the accuracy of PCA is discussed. Revised evaluation criteria for a scree plot are proposed. Nullspace based denoising (NBD) of large matrices is introduced.