Skip to main content

Table 1 Performance comparisons between XGBoost and Random Forrest on the pan-cancer dataset

From: Putative biomarkers for predicting tumor sample purity based on gene expression data

Method

Number of genes used

RMSE

Correlation coefficient

Pearson

Spearman

Random Forrest

all

0.151

0.940

0.780

XGBoost

all

0.124

0.806

0.813

5

0.150

0.701

0.708

10

0.146

0.719

0.726

20

0.139

0.751

0.756

100

0.130

0.787

0.794

200

0.128

0.792

0.799

300

0.126

0.800

0.807