cross.val: cross.val Runs k-fold cross validation via rolling window....

In abarciauskas-bgse/cphtbo: Mashable Popularity Predictions

Description

cross.val Runs k-fold cross validation via rolling window. Any required libraries to run the classification algorithm should already be loaded.

Usage

cross.val(model.function, model.args = list(), data.train, no.subsets = 5)

Arguments

model.function	function name of the classification algorithm call
model.args	arguments to pass to the function call
data.train	data frame or matrix of training data, depending on what the algorithm requires
no.subsets	number of folds to run cross-validation on. Defaults to 5.

Value

list(rates, mean.rates) - percentage successful predictions and mean success rate of training and testing no.subsets

Examples

# Example with multinom
# cross.val(model.function = multinom, model.args = list(formula = popularity ~ ., data = training), data = training)
# $rates
# [1] 0.4941667 0.4810417 0.4658333 0.4770833 0.4935417
# $mean.rate
# [1] 0.4823333

abarciauskas-bgse/cphtbo documentation built on May 10, 2019, 4:09 a.m.