Prediction intervals provide a measure of uncertainty for predictions on regression problems (for example, a 95% prediction interval indicates that 95 out of 100 times, the true value will fall between the lower and upper values of the range. Thus, this is different from a simple point prediction that might represent the center of the uncertainty interval.
How to write prediction intervals for the SARIMA hybridization code in R for univariate time series forecasting?