Introduction
The forecast package provides functions for time series forecasting. It implements various methods.
Basic Forecasting
library(forecast)
# Simple forecast
fit <- meanf(ts_data, h = 10)
# Naive
fit <- naive(ts_data, h = 10)
# Seasonal naive
fit <- snaive(ts_data, h = 10)
ARIMA
# Auto ARIMA
fit <- auto.arima(ts_data)
# Manual ARIMA
fit <- arima(ts_data, order = c(1, 1, 1))
# Forecast
forecast(fit, h = 10)
Exponential Smoothing
# Simple exponential smoothing
fit <- HoltWinters(ts_data, beta = FALSE, gamma = FALSE)
# Holt's method
fit <- HoltWinters(ts_data, gamma = FALSE)
# Forecast
forecast(fit, h = 10)
Plotting
autoplot(forecast(fit))
Summary
forecast provides forecasting methods. Use auto.arima for automatic model selection.