A comparison of ARIMA forecasting and heuristic modelling

• Autores: Wang Chi-Chen, Yun-Shen Hsu, Cheng-Hwai Liou
• Localización: Applied financial economics, ISSN 0960-3107, Vol. 21, Nº. 13-15, 2011, págs. 1095-1102
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • The study compares the application of the forecasting methods Autoregressive Integrated Moving Average (ARIMA) time series model and fuzzy time series by heuristic models on the amount of Taiwan export. When our model prolongs the sample period, the predicted error is smaller for the ARIMA model than for the heuristic model. Moreover, the predicted trajectory of the ARIMA model is much closer to the realistic trend than the heuristic model. Thus, the ARIMA model can forecast the export amount more accurately than the heuristic models. In the economic viewpoints, the amount of Taiwan exports is mainly attributable to external factors. In addition, the impact reduces with time and the export with lags 12 or 13 do not affect current export amount anymore. If the sample period is shorter, the heuristic models outperform ARIMA models. A heuristic fuzzy time series model can be utilized to predict export values accurately, when only small set of data is available.