Three Survival-Related Genes of Esophageal Squamous Cell Carcinoma Identified by Weighted Gene Coexpression Network Analysis

Figure 6

LASSO Cox regression model construction. (a) λ selection by 10-fold cross-validation. Continuous upright lines are partial likelihood deviance ± standard error (SE); imaginary lines are depicted at the optimal values by minimum criteria (lambda.min, left vertical dotted line) and 1−SE criteria (lambda.1se, right vertical dotted line). The partial likelihood deviance with changing of log (λ) is plotted. The value 0.073 is chosen for λ by 10-fold cross-validation with the minimum criteria. (b) Process of LASSO Cox model fitting. Each curve represents a gene. The trend of each coefficient against the L1-norm is plotted when λ changes. L1-norm is the total absolute of nonzero coefficients. LASSO: least absolute shrinkage and selection operator; L1-norm: L1 regularization, the total absolute of nonzero coefficients; SE: standard error.

(a)

(b)