Impact of decomposition on time series bagging forecasting performance

Time series bagging has been deemed an effective way to improve unstable modelling procedures and subsequent forecasting accuracy. However, the literature paid little attention decomposition in time bagging. This study investigates impacts of various methods on performance. Eight popular approaches are incorporated into procedure procedures, resulting methods' performance is evaluated. Using world's top 20 inbound destinations as empirical case, this generates one-to eight-step-ahead tourism forecasts compares them against benchmarks, including non-bagged seasonal naïve models. For short-term forecasts, constructed via extraction autoregressive integrated moving average outperforms other methods. An autocorrelation test shows that efficient reduces variance forecasts.