Random forest regression provides enhanced forecasts compared with other time series models, for example AR and ARIMA. Besides that, the Random forest model allows to work with different problems that classical time series models like AR, ARIMA and VAR can’t perform. For example, when you have a multivariate input for a univariate output.
Check out the full blog post on Medium.