Identifying Interactions Between Binary Predictors
Get the best number of boosting iterations
Fast computation of the AUC w.r.t. to the ROC
Calculate the Brier score
Calculate the deviance
Calculate the misclassification rate
Calculate the MSE
Calculate the normalized cross entropy
Calculate the NRMSE
Define the cooling schedule for simulated annealing
Optimal pruning via cross-validation
Fitting 4pL models
Linear models based on boosted models
Linear models based on logic terms
Fitting linear models
Tuning the LASSO regularization parameter
Design matrix for the set of conjunctions
Gene-environment (GxE) interaction test based on boosted linear models
Gene-environment interaction test
Term importance test based on boosted linear models
Fitting bagged logicDT models
Fitting boosted logicDT models
Fitting logic decision trees
Partial prediction for boosted models
Plot a logic decision tree
Plot calculated VIMs
Prediction for 4pL models
Prediction for linear.logic models
Prediction for linear models
Prediction for logicDT models
Pruning path of a logic decision tree
Post-pruning using a fixed complexity penalty
Refit the logic decision trees
Split biallelic SNPs into binary variables
Control parameters for fitting decision trees
Variable Importance Measures (VIMs)
A statistical learning method that tries to find the best set of predictors and interactions between predictors for modeling binary or quantitative response data in a decision tree. Several search algorithms and ensembling techniques are implemented allowing for finetuning the method to the specific problem. Interactions with quantitative covariables can be properly taken into account by fitting local regression models. Moreover, a variable importance measure for assessing marginal and interaction effects is provided. Implements the procedures proposed by Lau et al. (2024, <doi:10.1007/s10994-023-06488-6>).