混沌理論與類神經網路應用於亞洲單一貨幣-以市場情緒指標為例

亞洲各國經過1997年金融風暴與2008年金融海嘯，有共識使用單一貨幣，以穩定金融市場及加強亞洲地區的經濟體質。亞洲國家大多以出口為主，匯率的變動亦影響到企業與投資人的獲利。過去的研究大部分以總體經濟面觀察匯率的走勢，本研究試著以VIX 指數、Put/call ratio、CRB指數、黃金價格/黃金股價指數、石油價格等五種市場情緒指標檢測亞洲單一貨幣的走勢，以期能協助匯率預測。本研究利用三種方法檢驗(BDS分析、R/S分析和相關維度分析法)亞洲單一貨幣匯率是否具有混沌現象。預測方面，使用適合非線性預測的倒傳遞類神經模型及時間遞延類神經模型，並加入VIX 指數、Put/call ratio、CRB指數、黃金價格/黃金股價指數、石油價格等五項自變數進行模型的預測效果比較，研究樣本為不同加權方式計算而成的亞洲單一貨幣ACU1(SDR之計算), ACU2(SDR修正計算), CCU1(SDR計算), CCU2(SDR修正計算),和其他的單一貨幣，例如：亞元指數(ADXY)和Asian Monetary Unit (AMU)，研究期間為1992年3月至2011年6月。亞洲單一貨幣樣本經由BDS分析、 R/S分析，和相關維度分析測試，具有混沌現象，結果顯示亞洲單一貨幣為非線性的時間序列，具有混沌現象。類神經網路預測結果顯示，市場情緒指標包括VIX 指數、Put/call ratio、CRB指數、黃金價格/黃金股價指數、石油價格適用於亞洲單一貨幣的匯率未來走勢預測，倒傳遞類神經網路進行亞洲單一貨幣的預測效果較佳，均可提供投資者未來的投資資訊和評估決策。

關鍵字

倒傳遞類神經；時間遞延類神經；混沌現象；亞洲單一貨幣

並列摘要

Through the financial tsunami in 1997 and 2008, Asian states have a consensus to use the single currency unit against external attacks and enhance regional economic sound. Asian countries are focus on trade, so the impact of profitability for enterprises and investors were based on exchange rate movement. In the past, most of study observed the effect of the exchange rate movement linked with macroeconomic factors. This study used five market sentiment indicators (such as VIX index, Put/Call ratio, CRB index, Gold price/HUI, and oil price) to examine the trend of Asian single currency units, and expected to enhance the exchange rate forecast. This study examine that predict the trend of Asian Currency units by market sentiment indicators. Utilizing the three approaches (such as BDS test, R/S Analysis, and Correlation Dimension Analysis) this paper examined whether Asian single Currency units have the chaos phenomenon. This paper uses a suitable nonlinear prediction of Back-Propagation Neural Network (BPNN) and Time-Delay Recurrent Neural Network models (TDNN), joining five sentiment indicators to forecast and compared the performance of the models. The study samples are constructed a different weighting for single currency unit in Asian, including ACU1 [calculated by Special Drawing Rights (SDR)], ACU2 (calculated by SDR), CCU1 (calculated by SDR), CCU2 (calculated by SDR modification), and other related single currencies like Asian Monetary Unit (AMU) and Asian Dollar Index (ADXY Index). The period was obtained from 1992/3 to 2011/6. Though BDS test, R/S Analysis, and Correlation Dimension Analysis, the results showed that Asian Single Currency Units samples are nonlinear time series and have chaos phenomenon. The results of neural network forecast indicated that the sentiment indicator like VIX index, Put/Call ratio, CRB index, Gold price/HUI, and Oil Spot Price can effectively apply predicting in Asian Single Currency Units trend in future. And Back-propagation Network (BPN) has better forecasting performance for Single Currency Units in Asia to provide investors a valuable information for investment and to assess decision making in the future.

並列關鍵字

Market Sentiment indicators ； Asian Currency units ； BPNN ； TDRNN ； Chaos Phenomenon

參考文獻

8. Bo, Wan Zhen (2010), The Oil prices can be an explained by exchange rate movements? Verification of the Emerging Asian Countries, Department of Economics, Soochow University, Vol. 69, P29-P46.

1. Azoff, E. M. (1994), Neural Network Time Series Forecasting of Financial Markets, John Wiley & Sons Inc., New York.

2. Apostolos, Serletis and Periklis, Gogas (1997), Chaos in East European Black Market Exchange Rates, Journal of Economics, Vol. 51, P359-P385.

3. Ashok, K. Nag and Amit, Mitra (2002), Forecasting Daily Foreign Exchange Rates Using Genetically Optimized Neural Networks, Journal of Forecasting, Vol. 21, P501-P511.

4. Atin, Das and Pritha, Das (2007), Chaotic Analysis of the Foreign Exchange Rates, Journal of Mathematics and Computation, Vol. 185, P388-P396.

國際替代計量

混沌理論與類神經網路應用於亞洲單一貨幣-以市場情緒指標為例

未授權

主題瀏覽