Dagliyan, O., F. Uney-Yuksektepe, I.H. Kavakli and M. Turkay, Optimization Based Tumor Classification from Microarray Gene Expression Data, PLoS ONE, 6(2), e14579 (2011).
Before the run, make sure that you have GAMS&MATLAB interface. To install the
interface please
read "doc.pdf" in the "matgams.zip" file. Also, GAMS and CPLEX licences are
needed to run the code.
To run the code:
Open "driver_nfold.m" for n-fold-cross-validation or "driver_testset.m" for
test-set. Then just click on
"run". The percentage accuracy will be on the command window of MATLAB.
In driver_nfold.m you can change the fold by changing the variable “fold=10”. On
test set, you
should enter the number of samples in training and test set. Please delete the
generated files after
you run "driver_testset.m".
This matlab("driver_nfold.m" or "driver_testset.m") file takes the "input.xls"
as input and prepares
input files of GAMS scripts, and then calls gams scripts.
Input format should be the same as is in "input.xls".
Please look at "TestResult_New.txt" for more detailed results such as predicted
classes of each
sample and corresponding boxes.
There are 16 intersection elimination scripts(all are same); however, you dont
need to use all of them.
The number of intersection elimination scripts that you should use depend on the
complexity of the
data. If you dont change anything, the code will use 16 of them, though some may
be unnecessary.
However the result wont change.
Optimization scripts:
For all data sets: Supplementary.rar
For individual data sets: