- 學位論文
機器學習於外匯趨勢預測與交易策略之應用
Foreign exchange rate trend forecasting and trading rules using machine learning
摘要
外匯市場為最活躍的金融市場，每日的成交量非常龐大。外匯的波動深深地影響我們的生活，不論是對個人，或是對企業家，甚至影響了國家做經濟上的決策，因此，外匯的預測相當重要。希望能透過機器學習的技術，捕捉外匯市場的特性，做到更貼近的市場趨勢預估。 此篇研究選用的機器學習模型為隨機森林、xgboost演算法與長短期記憶模型，並藉此深入探討不同特徵對模型匯率趨勢預測的影響，如技術指標與總體經濟因子，與探究有無加入總體經濟因子對模型的影響程度。研究中，將探討如何利用機器學習的模型，來預測不同天數的外匯匯率市場趨勢，並採用相對應的交易策略，達到期望的投資績效。
並列摘要
The foreign exchange market is the most active financial market, and the daily trading volume is very large. The volatility of foreign exchange deeply affects our lives, whether it is for individuals, entrepreneurs, or even the government to make decisions. Therefore, forecasting foreign exchange is very important. We want to use machine learning technologies to capture the characteristics of the foreign exchange market and make predictions of the market trends. The machine learning models used in this study are random forests, xgboost algorithm, and long short-term memory model, and use the model to explore the impact of different features on the model, such as technical indicators and economic factors, and explore the influence of economic factors on the model's prediction. In the study, we will explore how to use machine learning models to predict the market trend of foreign exchange rates on different days, and use corresponding trading strategies to achieve desired investment performance.