LMLogisticLossEarlyStoppingCV: Cross validation algorithm using linear model with logistic...

In SixianZhang/CS499-Coding-Project-2: Linear model for regression and binary classification

Description

Training by using cross validation on a linear model with logistic loss and early stopping method. Return a list which contains the best iteration step, mean loss of training and validation data, and a predict function which gives a prediction based on the selected step.

Usage

LMLogisticLossEarlyStoppingCV(X.mat, y.vec, fold.vec = NULL,
  max.iteration, step.size = 0.5)

Arguments

X.mat	train feature matrix of size [n x p]
y.vec	train label vector of size [n x 1]
fold.vec	fold index vector of size [n x 1]
max.iteration	integer scalar greater than 1
step.size	a numeric scaler greater than 0, default is 0.5

Value

result.list a list with mean.validation.loss.vec, mean.train.loss.vec,selected.steps,weight.vec,and predict function

Examples

data(spam, package = "ElemStatLearn")
X.mat <- as.matrix(spam[, 1:57])
y.vec <- ifelse(spam$spam == "spam", 1, 0)
earlystopping.list <- LMLogisticLossEarlyStoppingCV(X.mat, y.vec, NULL, 100L, 0.5)
(earlystopping.list$predict(as.matrix(X.mat[c(2,4),])))

SixianZhang/CS499-Coding-Project-2 documentation built on May 26, 2019, 3:31 p.m.