nn_embedded_lstm: Neural Network with a long-term short-term memory and an...

In pchest/simpleNN: Provides Accessable Deep Learning Tools

Description

This function is a wrapper for a long term short term neural network written using the Keras Package.

Usage

nn_embedded_lstm(Text, Codes, Words = 3000, Seed = 17,
  Train_prop = 0.5, Epochs = 10, Batch = 32, MaxSentencelen = 60,
  WordEmbedDim = 50, ValSplit = 0.1, Units_lstm = 64,
  Dropout = 0.2, Recurrent_dropout = 0.2, CM = TRUE, Model = FALSE)

Arguments

Text	The text that will be used as training and test data.
Codes	The codes that will be used as outcomes to be predicted by the NN model.
Words	The number of top words included in document feature matrixes used as training and testing data.
Seed	The seed used in the model. Defaults to 17
Train_prop	The proportion of the data used to train the model. The remainder is used as test data.
Epochs	The number of epochs used in the NN model.
Batch	The number of batches estimated in the NN.
MaxSentencelen	All sentences will be truncated to this length to be input into the LSTM model
WordEmbedDim	The number of word embedding dimensions to be produced by the LSTM model
ValSplit	The validation split of the data used in the training of the LSTM model
Units_lstm	The number of network nodes used in the LSTM layer
Dropout	A floating variable bound between 0 and 1. It determines the rate at which units are dropped for the linear tranformation of the inputs.
Recurrent_dropout	A floating variable bound between 0 and 1. It determines the fraction of the units to drop for the linear transformation of the recurrent neural network layer.
CM	A logical variable that indicates whether a confusion matrix will be output from the function
Model	A logical variable that indicates whether the trained model should be included in the output of this function

pchest/simpleNN documentation built on May 14, 2019, 8:50 p.m.