fda_learn

This is a list of relevant scripts and functions produced while learning FDA (functional data analysis).

Contents

smooth_derivative.m: wrap up of the script for smoothing spline by GCV in:

Ramsay, James & Hooker, Giles & Graves, Spencer. (2009). Functional data analysis with R and MATLAB. 10.1007/978-0-387-98185-7.

fda_pca.m: wrap up of the script for fda pca for a fixed lambda

*_test.m: corresponding test script

Usage

1. Download and add the FDA matlab script (http://www.psych.mcgill.ca/misc/fda/downloads/FDAfuns/)
2. Try run smooth_derivative_test.m, especially the first example.
3. Run similar script on your own data. Each column of Ymat is one curve, step size in xvec need to be close to 1, and choose nDer according to your need. You might also need to try different range of loglambda_vec.

For resources on FDA, refer to:

http://www.psych.mcgill.ca/misc/fda/

https://cran.r-project.org/web/packages/fda/index.html

http://www.psych.mcgill.ca/misc/fda/downloads/FDAfuns/