Basic Study
Copyright ©The Author(s) 2019.
World J Gastrointest Oncol. Nov 15, 2019; 11(11): 983-997
Published online Nov 15, 2019. doi: 10.4251/wjgo.v11.i11.983
Figure 5
Figure 5 Conditional interference tree for prediction. The nodes of the conditional inference tree were calculated automatically based on the influence power of variables. The first node of the tree was risk_score, while the second was stage. Node 3, 4, or 5 was the corresponding group a certain patient would be assigned to. For example, a stage I patient with a risk_score ≤ 0.855 would be divided into group node 3, which was related to the best overall survival.