Table 5.
Coefficients, tuning parameters, and classification error rate for the two LASSO logistic models for pathological response (MD Anderson criteria) (26).
| Coefficients | without | with |
|---|---|---|
| PET response | PET response | |
| as covariate | as covariate | |
| Intercept | 0.405 | −1.842 |
| PET response | – | 1.313 |
| ID_Range | – | 0.006 |
| ID_CoefficientOfVariation | – | 0.001 |
| λmin | 0.332 | 0.204 |
| Classification Error rate | 0.400 | 0.133 |
The PET response is added to the second model as covariate. Both models allow the algorithm to select the covariates to be included in the model without constraints.