A regression-style function call for keras_model_sequential() which uses formulas and sparse matrices. A sequential model is a linear stack of layers.
kms(input_formula, data, keras_model_seq = NULL, layers = list(units =
c(256, 128, NA), activation = c("relu", "relu", "softmax"), dropout = c(0.4,
0.3, NA), use_bias = TRUE, kernel_initializer = NULL, kernel_regularizer =
"regularizer_l1", bias_regularizer = "regularizer_l1", activity_regularizer =
"regularizer_l1"), pTraining = 0.8, validation_split = 0.2,
Nepochs = 15, batch_size = 32, loss = NULL, metrics = NULL,
optimizer = "optimizer_adam", scale_continuous = "zero_one",
drop_intercept = TRUE, sparse_data = FALSE, seed = list(seed = NULL,
disable_gpu = FALSE, disable_parallel_cpu = FALSE), verbose = 1, ...)an object of class "formula" (or one coerceable to a formula): a symbolic description of the keras inputs. "mpg ~ cylinders". kms treats numeric data with more than two distinct values a continuous outcome for which a regression-style model is fit. Factors and character variables are classified; to force classification, "as.factor(cyl) ~ .".
a data.frame.
A compiled Keras sequential model. If non-NULL (NULL is the default), then bypasses the following `kms` parameters: layers, loss, metrics, and optimizer.
a list that creates a dense Keras model. Contains the number of units, activation type, and dropout rate. For classification, defaults to three layers: layers = list(units = c(256, 128, NA), activation = c("relu", "relu", "softmax"), dropout = c(0.4, 0.3, NA), ...). If the final element of units is NA (default), set to the number of unique elements in y. kms defines the number of layers as the length of the vector of activations. Inputs that appear once are repeated Nlayer times. See ?layer_dense or ?layer_dropout. For regression, activation = c("relu", "softmax", "linear"). For penalty terms, options must be precisely either "regularizer_l1", "regularizer_l2", or "regulizer_l1_l2". Also, "kernel_initalizer" defaults to "glorot_uniform" for classification and "glorot_normal" for regression (but either can be inputted with quotes).
Proportion of the data to be used for training the model; 0 =< pTraining < 1. By default, pTraining == 0.8. Other observations used only postestimation (e.g., confusion matrix).
Portion of data to be used for validating each epoch (i.e., portion of pTraining). To be passed to keras::fit. Default == 0.2.
Number of epochs; default == 15. To be passed to keras::fit.
To be passed to keras::fit and keras::predict_classes. Default == 32.
To be passed to keras::compile. Defaults to "binary_crossentropy", "categorical_crossentropy", or "mean_squared_error" based on input_formula and data.
Additional metric(s) beyond the loss function to be passed to keras::compile. Defaults to "mean_absolute_error" and "mean_absolute_percentage_error" for continuous and c("accuracy") for binary/categorical (as well whether whether examples are correctly classified in one of the top five most popular categories or not if the number of categories K > 20).
To be passed to keras::compile. Defaults to "optimizer_adam", an algorithm for first-order gradient-based optimization of stochastic objective functions introduced by Kingma and Ba (2015) here: https://arxiv.org/pdf/1412.6980v8.pdf.
Function to scale each non-binary column of the training data (and, if y is continuous, the outcome). The default 'scale_continuous = zero_one' places each non-binary column of the training model matrix on [0, 1]; 'scale_continuous = z' standardizes; 'scale_continuous = NULL' leaves the data on its original scale.
TRUE by default.
Default == FALSE. If TRUE, X is constructed by sparse.model.matrix() instead of model.matrix(). Recommended to improve memory usage if there are a large number of categorical variables or a few categorical variables with a large number of levels. May compromise speed, particularly if X is mostly numeric.
Integer or list containing seed to be passed to the sources of variation: R, Python's Numpy, and Tensorflow. If seed is NULL, automatically generated. Note setting seed ensures data will be partitioned in the same way but to ensure identical results, set disable_gpu = TRUE and disable_parallel_cpu = TRUE. Wrapper for use_session_with_seed(), which is to be called before compiling by the user if a compiled Keras model is passed into kms. See also see https://stackoverflow.com/questions/42022950/.
Default == 1. Setting to 0 disables progress bar and epoch-by-epoch plots (disabling them is recommended for knitting RMarkdowns if X11 not installed).
Additional parameters to be passsed to Matrix::sparse.model.matrix.
kms_fit object. A list containing model, predictions, evaluations, as well as other details like how the data were split into testing and training. To extract or save weights, see https://tensorflow.rstudio.com/keras/reference/save_model_hdf5.html
# NOT RUN {
if(is_keras_available()){
mtcars$make <- unlist(lapply(strsplit(rownames(mtcars), " "), function(tokens) tokens[1]))
company <- kms(make ~ ., mtcars, Nepochs = 1)
# out of sample accuracy
pCorrect <- mean(company$y_test == company$predictions)
pCorrect
company$confusion
# plot(history$company) # helps pick Nepochs
company <- kms(make ~ ., mtcars, Nepochs = 1, seed = 2018,
layers = list(units = c(11, 9, NA), activation = c("relu", "relu", "softmax"),
dropout = c(0.4, 0.3, NA)))
# ?predict.kms_fit to see how to predict on newdata
}else{
cat("Please run install_keras() before using kms(). ?install_keras for options like gpu.")
}
# }
Run the code above in your browser using DataLab