TimeGPT is a generative pre-trained transformer model for time series
analysis (https://docs.nixtla.io/docs) developed
by Nixtla (https://nixtla.io). It was trained on the largest
collection of public time series and can predict the future values of a
single time series based on the provided. This R package provides an R
interface to interact with the TimeGPT API.
You can install the development version of TimeGPTapi from GitHub with:
# install.packages("devtools")
devtools::install_github("danigiro/TimeGPTapi")This is a basic example which shows you how to solve a common problem:
library(TimeGPTapi)
# Loading the air passengers dataset and the time vector
y <- AirPassengers
time_air <- zoo::as.Date(AirPassengers)
# Set the TimeGPT API token
set_TimeGPT_token("<TIMEGPT_TOKEN>")
obj <- TimeGPT(AirPassengers, time_air, freq = "M", verbose = TRUE, historic = TRUE)
# Forecasting the next 12 horizons using TimeGPT
fore <- forecast(obj2, h = 12)- Implement external regressor (
xreg) - Multivariate TimeGPT
This R package is released under the GNU General Public License v3.0.
The TimeGPT model and API are proprietary technologies developed by
Nixtla, and all rights are reserved. The use of the TimeGPT API is
subject to Nixtla’s terms and conditions, which can be found at
https://nixtla.io.