在此篇論文中,我們實作了一個自動交易系統,結合交易策略評價的概念,可根據盤中即時資料自動進行模擬或實務操作。系統可擷取券商報價,自動進行以市場現況為準的模擬交易,代替人工的看盤及下單買賣動作,可做為期貨1999-2007年每日交易資料以及美國道瓊指數期貨1996-2007年之每日交易資料為模擬市場交易研究之技術平台。在策略評價部分,主要從獲利性方面探討,選擇台灣加權指數研究標的,依此進行數據分析。全球性股市脈動往往有其高度連動性,而其中美國道瓊指數又可說為最具指標特性。以此資料針對市場上數種知名指標進行獲利性研究,如MA、KD、RSI、MACD、William %R等。進而產生實務可行的交易策略,並依此交易法則建立自動化交易系統。
In this thesis, we implement an automatic trading system and evaluate some trading strategies. Simulated and practical operation can be executed automatically by real-time intraday data. System can capture transaction data provided by the market. It can also be used as a research platform. About the evaluation of strategies, we mainly focus on the profitability aspects and use the daily transaction data of Taiwan's Weighted Index futures from 1999 to 2007 and the Dow Jones index futures from 1996 to 2007. By back-testing of these data, we analyze several well-known trading strategies such as MA, KD, RSI, MACD, William %R, and so on.