In order to compare the performance of the sparse learning techniques, classical regression method including OLR and logistic regression with backward stepwise selection (LR-BS) were also constructed using the same training dataset.
The prediction model was developed with multivariable logistic regression with backward stepwise selection.
Multivariate analysis was performed using a multiple logistic regression model (backward stepwise selection).
To identify factors significantly associated with incidence of preterm birth and mortality, we initially included all covariates into the model and used logistic regression with backward stepwise selection.
A baseline clinical model for toxicity was determined using logistic regression and backward stepwise selection used to identify relevant clinical characteristics to be included in the model.
Multifactorial regression (i.e., regression with multiple independent variables but a single dependent variable) was based on linear or logistic regression with backward stepwise selection (P for removal 0.10).
Data were analyzed with multiple logistic regression and backward stepwise selection (maximum P = 0.01) with adjustment for important confounding variables, including age, sex, geography, insurance type, comorbidity scores, cardiovascular risk factors, diabetes complications, total baseline medical expenditures, and prior ACVEs.
†The LASSO and LR-BS models were trained in scenario 3. AUC Area under the receiver operating characteristic curve, CI Confidence interval, FPG Fasting plasma glucose, HbA1c Glycated hemoglobin, LASSO Least absolute shrinkage and selection operator, LR-BS Logistic regression with backward stepwise selection, NPV Negative predictive value, PPV Positive predictive value.
In a logistic regression by backward stepwise selection using biological factors only, the factors discriminating IBC from non-IBC were P-cadherin (OR = 4.9, p = 0.0019), MIB1 (OR = 3.6, p = 0.001), CK14 (OR = 2.7, p = 0.02), and ERBB2 (OR = 2.3, p = 0.06).
To test whether Id-1 and Id-2 might act as independent prognostic factors in our patient cohort, Cox-regression with backward stepwise selection was carried out.
