get_forecast_accuracy_overview: Get forecast accuracy overview.

In ing-bank/tsforecast: Time Series Forecasting Pipeline

Description

get_forecast_accuracy_overview is a function to create an overview of the accuracy of the different forecast models in the main forecasting table. The function combines the fc_errors data from all rows within the main_forecasting_table to calculate the overall accuracy of each forecast model for each number of periods ahead that has been forecasted.

Usage

get_forecast_accuracy_overview(
  main_forecasting_table,
  metric = c("MAE", "MAPE", "MASE")
)

Arguments

main_forecasting_table	A tibble containing several columns of data required for time series forecasting, which has been created using the create_main_forecasting_table function and which has been extended with the fc_models and fc_errors columns using the add_fc_models_to_main_forecasting_table function.
sort_by	Should the fc models be sored by the mean error values ("mean"), the median error values ("median"), the minimum error values ("min") or the maximum error values ("max").
abs_error	Boolean, which is to be set to TRUE if the absolute error values should be used to calculate the forecast accuracy, or set to FALSE if the error values as is should be used to calculate the forecast accuracy (which has the problem that positive and negative errors can cancel each other out, leading to non-intuitive conclusions about forecast accuracy).

Value

A tibble containing an overview of the overall forecast accuracy of each forecast model in the main_forecasting_table.

Examples

data <- tstools::initialize_ts_forecast_data(
   data = dummy_gasprice,
   date_col = "year_month",
   col_of_interest = "gasprice",
   group_cols = c("state", "oil_company")
)
main_forecasting_table <- create_main_forecasting_table(
      data = data,
      seasonal_periods = c(12,3)
   ) %>%
   add_fc_models_to_main_forecasting_table(
      periods_ahead = 12,
      fc_methods = c("basic", "linear")
   )
get_forecast_accuracy_overview(main_forecasting_table)

ing-bank/tsforecast documentation built on Sept. 18, 2020, 9:40 a.m.