- 期刊
隨機指標平滑曲線優化之投資績效分析－以元大寶來台灣卓越50指數股票型基金為例
Application of Smoothing Filters for Stochastic Oscillator in Investment - Evidence of Taiwan 50 ETF
摘要
本研究以元大寶來台灣卓越50指數股票型基金之週資料為樣本，利用Ehlers（2013）所提出的Noise Aliasing及Spectral Dilation之技術優化傳統隨機指標（stochastic oscillator）並衍生十二種交易模式波段頻率，進而比較優化前後之獲利績效。本研究之參數設定在9，並將進退場數值設定在20及80。研究結果發現，經優化後之十二種交易模式，其波段波幅會變大且較平滑，交易次數雖然明顯增多，然而卻改善了優化前部分區段無法有效進場操作甚至獲利之問題；另外，經優化後十二種交易模式之獲利績效均有明顯的改善，尤其是平均單筆虧損都有明顯的下降，其原因在優化後雖總交易次數增加，但單筆持有時間縮短有利於做空模式在多頭市場中停損，而在做多模式中又能有效利用其市場之優勢，其中又以當數值由上往下跌破20時買進，並在數值由上往下跌破80時賣出之獲利績效最好。
並列摘要
In this research, we applied the techniques of noise aliasing and spectral dilation (Ehlers,2013) to the weekly data sample of Taiwan 50 ETF from 2003 to 2020. Using 12 trading strategies based on pre- and post-noise aliasing techniques, the empirical results show that trading based on post-aliasing noise generate better profitability than that of pre-noise aliasing. In addition, despite the higher number of transactions, post-aliasing noise substantially improves the average losses per trade. Finally, among 12 trading strategies, the strategy of entering market when KD drops below 20 and exiting market when KD drops below 80 tend to deliver annualized rate of return higher than 9%.