Description Usage Arguments Value
gbm_baseline
This function builds a baseline model using gradient boosting machine algorithm.
1 2 3 4 5 6 | gbm_baseline(train_path = NULL, pred_path = NULL, days_off_path = NULL,
train_Data = NULL, pred_Data = NULL, k_folds = 5,
variables = c("Temp", "tow"), ncores = parallel::detectCores(logical = F),
cv_blocks = "weeks", iter = seq(from = 50, to = 300, by = 25),
depth = c(3:7), lr = c(0.05, 0.1), subsample = c(0.5),
verbose = FALSE)
|
train_path |
The path of the file from which the training data are to be read. |
pred_path |
The path of the file from which the prediction data are to be read. |
days_off_path |
The path of the file from which the date data of days off (e.g., holidays) are to be read. |
train_Data |
A dataframe, of the training period, where the columns correspond to the time steps (time), the energy load (eload) and to the Temperature (Temp). |
pred_Data |
A dataframe, of the prediction period, where the columns correspond to the time steps (time), the energy load (eload) and to the Temperature (Temp). |
k_folds |
An integer that corresponds to the number of CV folds. |
variables |
A vector that contains the names of the variables that will be considered by the function as input variables. |
ncores |
Number of threads used for the parallelization of the cross validation. |
cv_blocks |
Type of blocks for the cross validation; Default is "none", which correspond to the standard k-fold cross validation technique. |
iter |
A vector with combination of the number of iterations. |
depth |
A vector with combination of the maximum depths. |
lr |
A vector with combination of the learning rates. |
subsample |
A vector with combination of subsamples. |
a gbm_baseline object, which alist with the following components:
an object that has been created by the function xgboost, and which correspond to the optimal gbm model.
a dataframe that correspond to the training data after the cleaning and filtering function were applied.
the fitted values.
a dataframe that contains the goodness of fitting metrics.
a dataframe the training accuracy metrics (R2, RMSE and CVRMSE) and values of the tuning hype-parameters.
a list of the best hyper-parameters
a dataframe that correspond to the prediction data after the cleaning and filtering function were applied.
the predicted values.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.