gradDescent (version 3.0)

ADADELTA: ADADELTA Method Learning Function

Description

A function to build prediction model using ADADELTA method.

Usage

ADADELTA(dataTrain, maxIter = 10, momentum = 0.9, seed = NULL)

Arguments

dataTrain

a data.frame that representing training data (\(m \times n\)), where \(m\) is the number of instances and \(n\) is the number of variables where the last column is the output variable. dataTrain must have at least two columns and ten rows of data that contain only numbers (integer or float).

maxIter

the maximal number of iterations.

momentum

a float value represent momentum give a constant speed to learning process.

seed

a integer value for static random. Default value is NULL, which means the function will not do static random.

Value

a vector matrix of theta (coefficient) for linear model.

Details

This function based on SGD with an optimization to create an adaptive learning rate by hessian approximation correction approach. Correction and has less computation load than ADAGRAD. This method create an exclusive learning rate and doesn't need alpha parameter, but uses momentum parameter same as MGD and AGD.

References

M. D. Zeiler Adadelta: An Adaptive Learning Rate Method, arXiv: 1212.5701v1, pp. 1-6 (2012)

See Also

ADAGRAD, RMSPROP, ADAM

Examples

# NOT RUN {
##################################
## Learning and Build Model with ADADELTA
## load R Package data
data(gradDescentRData)
## get z-factor data
dataSet <- gradDescentRData$CompressilbilityFactor
## split dataset
splitedDataSet <- splitData(dataSet)
## build model with ADADELTA
ADADELTAmodel <- ADADELTA(splitedDataSet$dataTrain)
#show result
print(ADADELTAmodel)

# }