PCA before mapstd or vise versa for classification?

1 次查看（过去 30 天）
Nisreen Jaradat
Nisreen Jaradat 2017-12-9
评论： Nisreen Jaradat 2017-12-9
I have data. I want to use PCA and also normalize it. What I did first was use mapstd then PCA. But I got some weird numbers. Should I mapstd it again? Or should I have waited to mapstd it in the first place?

请先登录，再回答此问题。

回答（1 个）

the cyclist
the cyclist 2017-12-9
If you want to normalize your data before doing PCA, you want to mapstd before (and only before) PCA. I'm going to guess at the error you may have made.
To normalize the data for PCA, you probably want to normalize the columns. The mapstd function normalizes the rows. Did you take that into account? You probably want
normed_data = mapstd(data')'
  1 个评论
Nisreen Jaradat
Nisreen Jaradat 2017-12-9
I did actually. :/ The numbers wound up becoming really variant after PCA. Like, from -3 to +7 for one variable.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Dimensionality Reduction and Feature Extraction 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by