Hodrick-Prescott濾波器的參數調整對移動平均技術分析之績效影響──以日頻外匯價格為例

To Adjust the Smoothing Parameter of Hodrick-Prescott Filter for Advancing the Directional Accuracy of the Traditional Moving Average Rule: Evidence on the Daily-Frequency Exchange Price

本研究採用Hodrick-Prescott濾波器優化傳統移動平均法，藉由HPF產生去除雜訊波動之價格趨勢，以降低雜訊對判斷買賣交易訊號的影響。檢視過去研究，HPF普遍應用於低頻率經濟數據分析，甚少研究將HPF應用在高頻率之交易資料，以做為建立操作交易策略之依據。此外，HPF建立的價格趨勢取決於平滑參數λ值之設定，由λ值決定資訊漏損及雜訊消除之扺換關係，進而影響HPF之操作績效。因此本研究將HPF應用於高頻率外匯價格之動能交易策略，以檢視HPF應用在高頻資料之妥適性，並藉由調整λ參數，找出適合運用於高頻率數據分析之λ值。

Using the Hodrick-Prescott filter, this study extracts the price trend from an observed price series for advancing the directional accuracy of the traditional moving average rule (MA) which usually is noised by the short-term component of price variation. Hodrick-Prescott filter have been widely-used to generate the business cycle using low-frequency economic data, however, less study examined its fitness for using high-frequency trading price. In addition, the performance of HPF is sensitive to the smoothing parameter of HPF’s λ. This study aims to detect the validity of HPF for using high-frequency exchange price. Secondly, an appropriate smoothing parameter of HPF’s λ is determined by investigating the performance of the high frequency momentum trading strategy.