IABC robotic evolutionary model for the foreign exchange rate prediction in Central America trading agreement events

摘要

Taiwanese economy is extremely export-oriented. However, Taiwan also starts to actively look for opportunities for participation, because of the slowdown of multinational liberalizations, and as a result of the advance of the regional economic integration trend. As an example, Taiwan vigorously participates in FTAs with countries in Asia and Central America. International economic conditions are significantly important to the international trade and the volume of export/import volume, which is affected by the foreign exchange rate. If contemporary researchers could take full advantage of the exchange rate forecasting, Taiwan could maximize its trade surplus, thus boosting the economic growth. Conventional foreign exchange rate forecasting is usually provided by analyzing many financial indices or with the time-series method. Our goal is to produce the foreign exchange rate forecasting result by the robotic way with an evolutionary computing method called Interactive Artificial Bee Colony algorithm. Based on the event study methodology, the selected agreements include four FTA that are ECFA, BIA, ASTEP and ANZTEC, and the observation period setting is 70 days of pre-event period and 70 days of post-event period. This paper uses time series model (GARCH, EGARCH) and Interactive artificial bee colony (IABC) to establish the exchange rate predicting models. In addition, we adopt Mean Absolutely Percentage Error (MAPE) to compare the accuracy of exchange rate prediction. There are many exchange rate predicting models and the most frequently one to conduct maybe the time series model. This research reveals that even the IABC is relatively new it is the model has the best predictive ability among all the models.