Fig. 3From: 18F-FDG PET/CT-based radiomics nomogram could predict bone marrow involvement in pediatric neuroblastomaRadiomics feature selection using the least absolute shrinkage and selection operator (LASSO) regression. A The tuning parameter lambda (λ) in the LASSO regression model was selected via five-fold cross-validation based on minimum criteria. The LASSO regression model shows the best predictive performance when the λ value was set as 0.027167 and log(λ) was − 3.605761, at which point 25 features were selected. B The dotted vertical line was plotted at the selected λ value, resulting in 25 non-zero-coefficient featuresBack to article page