baselines: Forecast with baseline models

In thfuchs/tsRNN: Time Series Forecasting using Recurrent Neural Networks by Keras

Description

Get forecast results for the following simple forecasting methods:

• Naive forecast (naive)

• Seasonal naive forecast

• Mean forecast (meanf)

• Simple exponential smoothing forecasts (ses)

• Exponential smoothing with Holt's trend

Usage

baselines(data, test_size = NULL, acc_measure)

Arguments

data	Univariate time series (ts object)
test_size	Numeric vector of length one specifying length of test set (along with forecasting horizon). NULL by default which yield 2 times the frequency of the time series
acc_measure	Accuracy indicator. Valid options are "ME", "RMSE", "MAE", "MPE", "MAPE", "MASE" and "ACF1"

Value

list of forecasting methods with data.frame for accuracy indicators

Examples

# Calculate MAE and RMSE of simple forecasting models for Apple FCF
apple <- tsRNN::ts_apple
baselines(apple, acc_measure = c("MAE", "RMSE"))

# Change forecast horizon to 4 years
baselines(apple, test_size = 16, acc_measure = c("MAE", "RMSE"))

thfuchs/tsRNN documentation built on April 17, 2021, 11:03 p.m.