To run the codes of "Global and Local Structure Preservation for Feature Selection",

(1) add all files into your currrent Matlab path;
(2) run the "Demo.m".

Please note that:
(1) The "Kmatrix" in supervised case is calculated via labels, while it is calculated via a Gaussian kernel in unsupervised case.
(2) The parameters of "\mu" and "\lambda" shoule be carefully tuned.

Please donot hesitate to contact us if you have any problem in running the codes. Email: 1022xinwang.liu@gmail.com and leiw@uow.edu.au

Please cite our work as:
@article{LiuWZYL14,
author = {Xinwang Liu and Lei Wang and Jian Zhang and Jianping Yin and Huan Liu},
title = {Global and Local Structure Preservation for Feature Selection},
journal = {{IEEE} Trans. Neural Netw. Learning Syst.},
volume = {25},
number = {6},
pages = {1083-1095},
year = {2014}
}