Supplemental Table 1: Comparison of different feature selection methods and 10-fold cross-validation performances of Top100, Top150, Top200, log2FC_2.5 and log2FC_3.
Name |
Description |
TOP100 |
The top 100 most variable genes across different tumor types selected as features |
TOP150 |
The top 150 most variable genes across different tumor types selected as features |
TOP200 |
The top 200 most variable genes across different tumor types selected as features |
log2FC_2.5 |
The different gene expression log2 fold change value above 2.5 across different tumor types selected as features |
log2FC_3 |
The different gene expression log2 fold change value above 3 across different tumor types selected as features |
mtry |
Random forest parameter: number of variables randomly sampled as candidates at each split |
ntree |
Random forest parameter: Number of trees to grow |
Accuracy |
The accuracy of overall classification |
Kappa |
The Cohen's kappa coefficient |
AccuracySD |
The standard deviation of accuracy |
KappaSD |
The standard deviation of kappa |