Trains a binary decision tree classifier using MediBoost Tree-Structured Boosting
mediboost(x, y, catPredictors = NULL, depthLimit = 8,
learningRate = 1, gamma = 0, update = "exponential",
min.update = ifelse(update == "polynomial", 10, 1000),
weights = NULL, autoweights = FALSE, min.hessian = 0.01,
min.membership = 0, steps.past.min.membership = 2,
save.rpart = FALSE, verbose = TRUE, trace = 1)Matrix / Data frame of features
Integer -1, 1: Vector of binary outcomes
Optional: Logical vector indicating categorical features
Integer: Maximum depth of tree to grow
learning rate for the Newton Raphson step that updates the
function values of the node for update = "exponential"
Float (0, 1): Accelaration factor
String: "Exponential" or "Polynomial"
Logical: If TRUE, print messages to output
Outcome must be factor with two levels, the first level is the 'positive' class
lambda <- gamma/(1 - gamma)
Valdes Gilmer, Luna Jose, Eaton Eric, Ungar Lyle, Simone Charles and Solberg Timothy. MediBoost: a Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine.