Abstract: Based on TEI@I methodology, this paper proposes a forecasting framework on air passenger demand. First, ensemble empirical mode decomposition (EEMD) is applied to decompose the original air passenger demand data into a number of relatively simple modes, reducing the complexity of the data. Second, the extracted modes are thoroughly analyzed to capture hidden data characteristics, including complexity, stationarity and long-range correlation properties. These characteristics are then used to determine appropriate forecasting models for each mode (econometric models or artificial intelligence models). After that, the impacts of irregular and the infrequent future factors on air passenger demand are explored using expert systems techniques. Finally, the components above are predicted independently and these prediction results are combined as an aggregated output. The empirical results indicate that the proposed model based on TEI@I methodology has a good prediction performance on air passenger demand.

Key words: TEI@I methodology, air passenger demand, detrended fluctuation analysis, sample entropy, long and short-term memory model