autoenc)

Description Usage Arguments Value

View source: R/autoenc.R

Implementation based on http://web.stanford.edu/class/archive/cs/cs294a/cs294a.1104/sparseAutoencoder.pdf

sparse_autoenc(X, num_hidden, activation, lambda, beta, rho, epsilon,
  tolerance = sqrt(.Machine$double.eps), X.test = NULL,
  d_activation = NULL, optim_method = "BFGS", max_iterations = 2000,
  rescale = TRUE, rescaling_offset = 0.001)

`X`	matrix. Training dataset.
`num_hidden`	integer. Specifies number of neurons in each hidden layer.
`activation`	character or function. If it's a character it has to be one of the predefined activation functions. If it's a function, `d_activation`, activation prime, function must be supplied as well. Those functions are needed for backpropagation.
`lambda`	numeric. Weight decay parameter.
`beta`	numeric. Learning rate parameter for algorithm trying to (approximately) satisfy the sparsity constraint.
`rho`	numeric. Sparsity parameter, which specifies our desired level of sparsity.
`epsilon`	numeric. A (small) parameter for initialization of weight matrices as small gaussian random numbers sampled from `N(0, epsilon^2)`
`tolerance`	numeric (optional). Tolarance to be used for comparing floting point numbers. Default is `.Machine$double.eps`
`X.test`	matrix (optional). Testing dataset for evaluating the network.
`d_activation`	function (optional). This parameter will be omitted if activation is a character, otherwise this function will be used in backpropagation as activation prime, in which case this parameter becomes mandatory.
`optim_method`	character (optional). Optimization method to be used. Please check `autoenc:::constants$optim_methods` for available options. Default is `"BFGS"`.
`max_iterations`	integer (optional). Maximum number of iterations for optimizer. Default is `2000`.
`rescale`	logical or numeric. Autoencoders yield better results when used on normalized matrices. Normalization should be performed according to the activation function being used. Default is `TRUE`, which will select the appropritate ranges for activation functions included in the package. If you are providing custom activation function, you must specify an appropriate range, like `c(-1, 1)`. If it is `FALSE` no rescaling will be performed.
`rescaling_offest`	numeric. A small value used in rescaling to `c(rescale[1] + offset, rescale[2] - offset)` interval. Default is `0.001`.