fitLSTMmodel: fitLSTMmodel

In MarkusBonsch/mxLSTM: LSTM model for regression

Description

Fits an mxNet LSTM recurrent neural network. For use in getLSTMmodel

Usage

fitLSTMmodel(x, y, param, test.x = NULL, test.y = NULL,
  initialModel = NULL, seed = NULL, ...)

Arguments

x	array containing the x variables for training. See mxLSTM for format.
y	array containing the y variables for training. See mxLSTM for format.
param	Parameters as given by caret's tuneGrid. See getLSTMmodel for details.
test.x	same as x, but for testing after each epoch
test.y	same as y, but for testin after each epoch
initialModel	A pretrained mxLSTM model as returned by this function. Training will continue on initialModel, using its weights as a starting point. Be careful, all data shape, hidden layers, and preProcessing must be identical to the first training of initialModel.
seed	Random seed tp be set before model training. Defaults to NULL for no explicit seed setting
...	further arguments for the call to mxLSTM

Value

list of class mxLSTM. This is the result of mxLSTM Additionally, some other list items are added:

• xVariables: Names of the xVariables

• yVariables: Names of the yVariables

See Also

mxLSTM, predictLSTMmodel, getLSTMmodel

MarkusBonsch/mxLSTM documentation built on May 28, 2019, 6:40 a.m.