batch_run_energy_retrofit: Batch run for specific energy (retrofit)

Description Usage Arguments Value Examples

View source: R/retrofit_controller.R

Description

This function run batches given given a utility data frame of retrofit data and energy. See retrofit_utility.

Usage

1
2
3
batch_run_energy_retrofit(utility, energy, metric_vec = TRUE,
  plot_flag = FALSE, step = 0.5, CV_RMSE_n = 0.25,
  Rsquared_n = 0.75, n_pre = 4, n_post = 4)

Arguments

utility

A utility data frame of retrofitted data with multiple buildings.

energy

A character string. Energy Type, either 'Elec' or 'Fuel'.

metric_vec

A numeric (integer) vector. This vector determines the pirority of model types. Defaults to c(1,2,3,4,5) for models: '2P', '3PC', '3PH', '4P', '5P'.

plot_flag

A boolean value. Defaults to FALSE. If set to FALSE, it will not return any plots.

step

A single numeric value or a vector. If it is a single numeric value, create_model will be used to find change-point(s). If it is a vector with more than one component, create_model_2 will be used. Defaults to 0.5. Adjust step-size depending on independent varaiables.

CV_RMSE_n

A numeric value. CV-RMSE threshold value. Defaults to 0.25.

Rsquared_n

A numeric value. Rsquared threshold value. Defaults to 0.75.

n_pre

A numeric value that determines threshold for population test for pre-retrofit: thereshold = number_of_independent_variables/n. Defaults to 4. See pop_test.

n_post

A numeric value that determines threshold for population test for post-retrofit: thereshold = number_of_independent_variables/n. Defaults to 4. See pop_test.

Value

A list with components:

best_result_df

A data frame of best models of mulitple buildings. If a building does not model, it will not be shown in this data frame.

plot_list

A list containing plots. If plot_flag is set to FALSE, this will be an empty list.

Examples

1
2
3
4
## Not run: 
batch_result = batch_run_energy_retrofit(retrofit_utility, 'Elec')

## End(Not run)

tinnaing347/bRema documentation built on Aug. 23, 2019, 9:15 p.m.