Hi Ahmed,

This is a great tutorial on KPCA: http://www.stat.yale.edu/~snn7/papers/kernelPCA.pdf

Best,

There are a wide range of tutorial articles on dimensionality reduction:

[1] "A survey of dimension reduction techniques" by Fodor.

[2] "Dimensionality Reduction: A Comparative Review" by Maten et al.

[3] "Dimension Reduction: A Guided Tour" by Burges.

The only one that describes all three methods is [1], while the other two focuse more on PCA and KPCA (between others).

Depending on the programming language you prefer, there also several existing implementations that you can study to understand these techniques better, e.g. the Matlab Toolbox for Dimensionality Reduction has PCA and KPCA.

Preisendorfer book 'Principal Component Analysis' is my favourite together with 'Lewi's Multivariate data analysis in industrial practice, and this one:

A User's Guide to Principal Components by J. Edward Jackson

you can find on the web

Dear Ahmed, here attached a very interesting and vivid introductory 'tale'

Thanks all for your help

Http://www.cs.otago.ac.nz/cosc453/student_tutorials/principal_components.pdf

 I am sorry to  poke my nose here but I have a question. I have 2 years data with four seasons in each year and 9 sites having almost 11 variables.

I want to do PCA but I do not know how to arrange my data?

Can I do PCA on my whole data set?

or Should I do it by year wise?

or Season wise?

Thanks In advance

Yes, Riaz you can do it as a whole, but I see that you may arrange the data as seasons for the 2 years and consider each group as PC.

My suggestion: https://www.researchgate.net/publication/319469038_New_Interpretation_of_Principal_Components_Analysis

Article New Interpretation of Principal Components Analysis