The "forecastHybrid" package provides functions to build composite models using multiple individual component models from the "forecast" package. These hybridModel objects can then be manipulated with many of the familiar functions from the "forecast" and "stats" packages including forecast(), plot(), accuracy(), residuals(), and fitted().

Basic usage

First load the package.

library(forecastHybrid)

Quick start

If you don't have time to read the whole guide and want to get started immediately with sane default settings to forecast the USAccDeaths timeseries, run the following:

quickModel <- hybridModel(USAccDeaths)

## Fitting the auto.arima model

## Fitting the ets model

## Fitting the thetam model

## Fitting the nnetar model

## Fitting the stlm model

## Fitting the tbats model

forecast(quickModel)

##          Point Forecast    Lo 80     Hi 80    Lo 95     Hi 95
## Jan 1979       8354.580 7924.712  8968.899 7706.957  9272.235
## Feb 1979       7543.318 6864.957  8184.977 6542.338  8468.428
## Mar 1979       8241.755 7223.731  8886.679 6888.559  9146.115
## Apr 1979       8531.595 7606.103  9194.629 7249.674  9500.477
## May 1979       9336.334 8105.778 10112.349 7734.506 10442.376
## Jun 1979       9776.752 8519.977 10525.745 8226.434 10878.297
## Jul 1979      10683.548 9158.635 11613.448 8784.188 11987.171
## Aug 1979       9986.536 8979.890 10830.331 8697.308 11224.086
## Sep 1979       9001.395 8281.314  9944.791 7892.801 10357.609
## Oct 1979       9256.502 8309.414 10198.510 8056.898 10629.548
## Nov 1979       8781.928 8033.333  9732.246 7589.183 10180.765
## Dec 1979       9112.847 8255.912 10255.628 7816.759 10720.971
## Jan 1980       8382.947 7485.300  9496.297 7033.420 10011.749
## Feb 1980       7600.342 6687.306  8749.398 6141.503  9295.201
## Mar 1980       8249.630 7033.489  9586.522 6637.427 10161.075
## Apr 1980       8542.046 7338.446  9940.466 7010.225 10542.396
## May 1980       9333.203 7749.995 10861.976 7355.042 11490.092
## Jun 1980       9762.515 8118.797 11280.306 7714.097 11933.559
## Jul 1980      10654.632 8600.469 12373.743 8150.610 13051.201
## Aug 1980       9963.104 8598.231 11596.927 8248.308 12297.755
## Sep 1980       8988.101 7949.455 10718.099 7261.423 11441.542
## Oct 1980       9250.330 8063.717 10978.824 7417.369 11724.196
## Nov 1980       8788.215 7623.215 10519.779 6856.540 11286.453
## Dec 1980       9106.311 7938.375 11050.531 7288.262 11837.932

plot(forecast(quickModel), main = "Forecast from auto.arima, ets, thetam, nnetar, stlm, and tbats model")

Fitting a model

The workhorse function of the package is hybridModel(), a function that combines several component models from the "forecast" package. At a minimum, the user must supply a ts or numeric vector for y. In this case, the ensemble will include all six component models: auto.arima(), ets(), thetam(), nnetar(), stlm(), and tbats(). To instead use only a subset of these models, pass a character string to the models argument with the first letter of each model to include. For example, to build an ensemble model on a simulated dataset with auto.arima(), ets(), and tbats() components, run

# Build a hybrid forecast on a simulated dataset using auto.arima, ets, and tbats models.
# Each model is given equal weight 
set.seed(12345)
series <- ts(rnorm(18), f = 2)
hm1 <- hybridModel(y = series, models = "aet", weights = "equal")

## Fitting the auto.arima model

## Fitting the ets model

## Fitting the tbats model

The individual component models are stored inside the hybridModel objects and can viewed in their respective slots, and all the regular methods from the "forecast" package could be applied to these individual component models.

# View the individual models 
hm1$auto.arima

## Series: y 
## ARIMA(0,0,0) with zero mean 
## 
## sigma^2 estimated as 0.6659:  log likelihood=-21.88
## AIC=45.76   AICc=46.01   BIC=46.65

# See forecasts from the auto.arima model
plot(forecast(hm1$auto.arima))

Model diagnostics

The hybridModel() function produces an S3 object of class forecastHybrid.

class(hm1)

## [1] "hybridModel"

is.hybridModel(hm1)

## [1] TRUE

The print() and summary() methods print information about the ensemble model including the weights assigned to each individual component model.

print(hm1)

## Hybrid forecast model comprised of the following models: auto.arima, ets, tbats
## ############
## auto.arima with weight 0.333 
## ############
## ets with weight 0.333 
## ############
## tbats with weight 0.333

summary(hm1)

##            Length Class          Mode     
## auto.arima 18     forecast_ARIMA list     
## ets        19     ets            list     
## tbats      21     bats           list     
## weights     3     -none-         numeric  
## frequency   1     -none-         numeric  
## x          18     ts             numeric  
## xreg        1     -none-         list     
## models      3     -none-         character
## fitted     18     -none-         numeric  
## residuals  18     ts             numeric

Two types of plots can be created for the created ensemble model: either a plot showing the actual and fitted value of each component model on the data or individual plots of the component models as created by their regular S3 plot() methods. Note that a plot() method does not exist in the "forecast" package for objects generated with stlm(), so this component model will be ignored when type = "models", but the other component models will be plotted regardless.

plot(quickModel, type = "fit")

plot(quickModel, type = "models")

Since version 0.4.0, ggplot graphs are available. Note, however, that the nnetar, and tbats models do not have ggplot::autoplot() methods, so these are not plotted.

plot(quickModel, type = "fit", ggplot = TRUE)

## Warning: Removed 12 row(s) containing missing values (geom_path).

Using the "forecastHybrid" package

David Shaub

2020-08-27

Installation

Basic usage

Quick start

Fitting a model

Model diagnostics