Skip to main content

Table 2 Best combinations of feature selection and machine learning methods (highest mean CIs) for each strategy. Standard deviations of CIs are also provided here and in the supplementary Table 3. Slash “/”separated SDs related to their corresponding feature selections

From: Development and validation of survival prognostic models for head and neck cancer patients using machine learning and dosiomics and CT radiomics features: a multicentric study

ML model

Feature selection method

Highest CI ± SD

Strategy

Coxph

CI/ MD/ VH

0.72 ± 0.14 / 0.15 / 0.15

DOSE/LLRR/LLRR

CoxBoost

MD

0.73 ± 0.15

LLRR

glmnet

VH

0.73 ± 0.15

LLRR

RSF

MD

0.71 ± 0.11

DOSE

glmboost

MD/ VH

0.72 ± 0.15 / 0.14

LLRR