Forecasting with `ahead`

T. Moudiki

5 months ago

[This article was first published on T. Moudiki's Webpage - R, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

A few weeks ago, I introduced a Forecasting API that I deployed on Heroku. Under the hood, this API is built on top of ahead (and through Python packages rpy2 and Flask); an R package for univariate and multivariate time series forecasting. As of October 13th, 2021, 5 forecasting methods are implemented in ahead:

armagarchf: univariate time series forecasting method using simulation of an ARMA(1, 1) – GARCH(1, 1)
dynrmf: univariate time series forecasting method adapted from forecast::nnetar to support any Statistical/Machine learning model (such as Ridge Regression, Random Forest, Support Vector Machines, etc.)
eatf: univariate time series forecasting method based on combinations of forecast::ets, forecast::auto.arima, and forecast::thetaf
ridge2f: multivariate time series forecasting method, based on quasi-randomized networks and presented in this paper
varf: multivariate time series forecasting method using Vector AutoRegressive model (VAR, mostly here for benchmarking purpose)

Here’s how to install the package:

1st method: from R-universe

In R console:

  options(repos = c(
      techtonique = 'https://techtonique.r-universe.dev',
      CRAN = 'https://cloud.r-project.org'))
        
  install.packages("ahead")

2nd method: from Github

In R console:

  devtools::install_github("Techtonique/ahead")

  remotes::install_github("Techtonique/ahead")

And here are the packages that will be used for this demo:

library(ahead)
library(fpp)
library(datasets)
library(randomForest)
library(e1071)

Univariate time series

In this section, we illustrate dynrmf for univariate time series forecasting, using Random Forest and SVMs. Do not hesitate to type ?dynrmf, ?armagarchf or ?eatf in R console for more details and examples.

par(mfrow=c(2, 2))

# Plotting forecasts
# With a Random Forest regressor, an horizon of 20, 
# and a 95% prediction interval
plot(dynrmf(fdeaths, h=20, level=95, fit_func = randomForest::randomForest,
      fit_params = list(ntree = 50), predict_func = predict))

# With a Support Vector Machine regressor, an horizon of 20, 
# and a 95% prediction interval
plot(dynrmf(fdeaths, h=20, level=95, fit_func = e1071::svm,
fit_params = list(kernel = "linear"), predict_func = predict))

plot(dynrmf(Nile, h=20, level=95, fit_func = randomForest::randomForest,
      fit_params = list(ntree = 50), predict_func = predict))

plot(dynrmf(Nile, h=20, level=95, fit_func = e1071::svm,
fit_params = list(kernel = "linear"), predict_func = predict))

Multivariate time series

In this section, we illustrate ridge2f and varf forecasting for multivariate time series. Do not hesitate to type ?ridge2f or ?varf in R console for more details on both functions.

# Forecast using ridge2
# With 2 time series lags, an horizon of 10, 
# and a 95% prediction interval
 fit_obj_ridge2 <- ahead::ridge2f(fpp::insurance, lags = 2,
                                  h = 10, level = 95)


# Forecast using VAR
 fit_obj_VAR <- ahead::varf(fpp::insurance, lags = 2,
                            h = 10, level = 95)

 
# Plotting forecasts 
# fpp::insurance contains 2 time series, Quotes and TV.advert 
 par(mfrow=c(2, 2))
 plot(fit_obj_ridge2, "Quotes")
 plot(fit_obj_VAR, "Quotes")
 plot(fit_obj_ridge2, "TV.advert")
 plot(fit_obj_VAR, "TV.advert")

To leave a comment for the author, please follow the link and comment on their blog: T. Moudiki's Webpage - R.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.