摘要
This study utilized two forecasting methods including ARFIMA （p, d, q）-GARCH （p, d, q）, and extreme value techniques. One of the puzzling questions raised by evolutionary econometric theory is how two-way behavior is evaluated in two ways, which benefits the investors of the securities, traded on a stock exchange. For the purpose of this study, intra-day secondary data during period of 1997-2010 of the stock-market returns of Bangkok SET （Stock Exchange of Thailand） Index （Thailand） and Kuala Lumpur Composite Index （Malaysia） were collected. For the new perspective framework, the expected values were conducted using ARFIMA （p, d, q）-GARCH （p, q） forecasting method and Generalize Extreme Value （GEV） to confirm the final solutions. The Value-at-Risk （VaR） of those stock-market returns was tested. The new perspective framework of expected value confirmed that ARFIMA （1, 0.29, 1）-GARCH （1, 1） was the best forecasting method for VaR in case of the Kuala Lumpur Composite stock-market returns based on MAPE （%）. And the perspective based on extreme case confirmed that Generalize Extreme Value （GEV） as F= （x,μ,σ,ξ）： F = （x, 0.00616, 0.00573, 0.36900） was the best forecasting method for VaR in case of the Bangkok SET stock-market returns based on MAPE （%）.
