Environment:
Linux/Windows + Matlab

Run the code:
1. Extract the package to one directory. Suppose it is "/source".
2. Launch Matlab and change the current directory to "/source".
3. Execute the corresponding Matlab scripts to do the experiments in the paper.
test_Rosetta_231.m: Do classification on Rosetta dataset with 231 important genes.
test_Rosetta_top500.m: Do classification on Rosetta dataset with top 500 genes ranked by correlation coefficient.
test_Wang_top700.m: Do classification on Wang dataset with top 700 genes ranked by correlation coefficient.
4. The classification performance of hypergraph-based algorithm and HyperGene are saved in a vector variable ROC_Hyper1, in which the first element is the ROC score of hypergraph-based algorithm and the last element is the ROC score of HyperGene. The ROC score for linear SVM and RBF SVM are saved in ROC_SVM0 and ROC_SVM1 separately.

Descriptions:
Get_f_a.m: Calculate the coefficient of the linear term in quadratic programming.
Get_ROC.m: Calculate the ROC score.
HypergraphClassify.m: Classify test set with hypergraph-based algorithm and HyperGene.
NetworkProp.m: Implement the hypergraph-beased algorithm.
NetworkProp2.m: Implement the HyperGene algorithm.
OptimizeF.m: Implement the first optimization step. 
OptimizeW.m: Implement the second optimization step.
SVM_Classify.m: Implement the SVM algorithm.
Rosetta.mat: Rosetta dataset.
Rosetta_HPPI_1000.mat: Human protein-protein interaction network for top 1000 ranked genes in Rosetta dataset.
Rosetta_HPPI_231.mat: Human protein-protein interaction network for 231 important genes in Rosetta dataset.
Wang.mat: Wang dataset.
Wang_HPPI_1000.mat: Human protein-protein interaction network for top 1000 ranked genes in Wang dataset.