Research Article
Random Subspace Aggregation for Cancer Prediction with Gene Expression Profiles
| Input: | | Dataset , sample size ; | | Sample , number of total feature ; | | Class of th sample in normal, tumor}; | | Split function: yield training set and testing set from original dataset. If the original | | dataset has been divided into training and testing partition, this step could be skipped. | | Gene select function: , where is the feature number of selected data, ; | | RS_preject function: , where is the size of a random subspace, ; | | Number of random subspaces ; | | Learning algorithm: SVM | | Output: | | Classification hypotheses : | | Start: | | Data processing: | | (Trainset, Testset) = Split() | | TrainsetNew = Gene_select(Trainset, ) | | TestsetNew = Gene_select(Testset, ) | | Generate and aggregate SVM classifiers on random subspaces: | | For to | | _project(TrainsetNew, ) | | () | | End | | Test: | | For each in TestsetNew | | | | End | | End |
|