plr_bootstrap_output_from_results: Bootstrap: Resample from individual Models

In PVplr: Performance Loss Rate Analysis Pipeline

Bootstrap: Resample from individual Models

Description

The function samples and bootstraps data that has already been put through a power predictive model. The PLR and Uncertainty are returned in a dataframe.

Usage

plr_bootstrap_output_from_results(
  data,
  power_var,
  time_var,
  weight_var,
  by = "month",
  model,
  fraction = 0.65,
  n = 1000
)

Arguments

data	Result of modeling data with a PLR determining model, i.e. plr_xbx_model, plr_6k_model, etc.
power_var	Variable name of power in the dataframe. Typically power_var
time_var	Variable name of time in the dataframe. Typically time_var
weight_var	Variable name of weightings in the dataframe. Typically sigma
by	String, either "day", "month", or "year". Time over which to perform plr_yoy_regression and plr_weighted_regression.
model	The name of the model the data has been put through. This option is only included for the user's benefit in keeping bootstrap outputs consistent.
fraction	The fractional size of the data to be sampled each time.
n	The number of resamples to take.

Value

Returns PLR value and uncertainty calculated with bootstrap of data going into power correction models

Examples

# build var_list


var_list <- plr_build_var_list(time_var = "timestamp",
                               power_var = "power",
                               irrad_var = "g_poa",
                               temp_var = "mod_temp",
                               wind_var = NA)
# Clean Data
test_dfc <- plr_cleaning(test_df, var_list, irrad_thresh = 100,
                         low_power_thresh = 0.01, high_power_cutoff = NA)
                         
# Perform the power predictive modeling step
test_xbx_wbw_res <- plr_xbx_model(test_dfc, var_list, by = "week",
                                  data_cutoff = 30, predict_data = NULL)
                                  
xbx_mbm_plr_result_uncertainty <- plr_bootstrap_output_from_results(test_xbx_wbw_res, 
                                                                    power_var = 'power_var',
                                                                    time_var = 'time_var',
                                                                    weight_var = 'sigma',
                                                                    by = "month", model = 'xbx',
                                                                    fraction = 0.65, n = 10)

PVplr documentation built on Feb. 16, 2023, 9:56 p.m.