MOFA (version 1.3.1)

getDefaultTrainOptions: Get default training options

Description

Function to obtain the default training options.

Usage

getDefaultTrainOptions()

Value

Returns a list with default training options, which have to be passed as an argument to prepareMOFA

Details

The training options are the following:

• maxiter: numeric value indicating the maximum number of iterations. Default is 5000, but we recommend using the 'tolerance' as convergence criteria.

• tolerance: numeric value indicating the convergence threshold based on the change in Evidence Lower Bound (deltaELBO). For quick exploration we recommend this to be around 1.0, and for a thorough training we recommend a value of 0.01. Default is 0.1

• DropFactorThreshold: numeric hyperparamter to automatically learn the number of factors. It indicates the threshold on fraction of variance explained to consider a factor inactive and automatically drop it from the model during training. For example, a value of 0.01 implies that factors explaining less than 1% of variance (in each view) will be dropped. Default is 0, which implies that only factors that explain no variance at all will be removed

• verbose: logical indicating whether to generate a verbose output.

• seed: random seed for reproducibility (default is NULL, which samples a random seed).

Examples

# NOT RUN {
TrainOptions <- getDefaultTrainOptions()
TrainOptions
# }