LDA for high dimension small sample size data

Linear discriminant analysis when the data dimensionality is high and sample size is much smaller

For medical imaging/ biological data where typically data dimensionality is huge typical LDA applications fail on account of attempting to invert an impossibly large matrix. This can be remedied using the kernel trick. The current file attempts to do this. If you use this work as a part of your assignment/thesis/paper please cite the following paper:
Gaonkar, Bilwaj, and Christos Davatzikos. "Analytic estimation of statistical significance maps for support vector machine based multi-variate image analysis and classification." NeuroImage 78 (2013): 270-283.

Please note that this code was developed and distributed as a part of the work towards the above paper and citing this helps us develop and distribute more such work.

I have made an attempt to put in appropriate comments in the code and this code is inspired by several previous codes submitted on matlab central