標題: 應用市場自我產生之訊息於台灣衍生性金融商品最適化短線交易策略之研究
Applying Market-generated Information to Discover Optimal Strategy of Short-term Trading on Taiwan Derivatives
作者: 朱陳彬
Ju, Chern-Bin
陳安斌
Chen, An-Pin
資訊管理研究所
關鍵字: 市場輪廓;倒傳遞類神經網路;未平倉量;衍生性金融商品;Market;BPNN;ProfileOpen Interest;Derivatives
公開日期: 2012
摘要: 隨著電子下單的普及,市場自我產生之訊息已瞬息萬變,投資人若不藉由有效的方法加以分析,便被這海量的資訊所淹沒。市場輪廓理論對於市場自我產生之訊息能有效的發覺其中邏輯之變化,本研究即以市場輪廓理論為基礎,嘗試研究價值區間的改變、交易者的籌碼變化及開市價格偏離值,對台股未來走勢的影響。更進一步,希望能透過市場輪廓的分析,窺知台股未來走勢,並以此作為投資的依據。 本研究利用倒傳遞類神經網路強大的非線性搜尋能力,以價值區間為基礎,並搭配開市價格偏離值,以及機構投資人在期貨與選擇權的未平倉量消長變化,作為學習的因子,以針對未來台指期的漲跌趨勢做預測,並將預測結果應用於不同的投資策略。包含以台指期為標的物的投機交易;賣出勒式部位並進行避險的避險交易;賣出選擇權以獲取時間價值的消極交易。除此之外,為了探討市場輪廓理論最佳的預測時間長度,在投機交易中,本研究針對預測當日、未來三日、未來五日設計三個不同實驗組進行交易,期望藉由實驗的設計找出最佳的投資期間。 研究結果顯示,以市場輪廓理論預測台股走勢,在投機交易上,對於未來三日的漲跌進行投資,有較佳的獲利效果;在避險交易上,能有效降低最大虧損;在消極交易上,有較佳的準確率。投資人可依需求選擇適合的交易策略進行投資。
Market-generated information has exploded rapidly since the popularity of electronic orders. Investors need to find effective method to analyze it. Otherwise, they will be overwhelmed by lots of information. Market Profile can effectively find the logical rules of market-generated information. Therefore, this research intends to use Market Profile as basic theory, trying to learn the relationship between the trend of TAIEX index and Market Profile that includes Value Area changes, institutional investors’ holdings and opening types. Hopefully get the insightful view of future index trend and profit from the investment with the suggestion of the research result. This research employs back-propagation neutral network (BPNN) with its tremendous non-linear searching ability to learn the behaviors of Value area changes, opening types and how institutional investors arrange their positions on futures and options. By input these variables to BPNN, to get predictive trend of index on Taiwan futures. In addition, this research will use the predictive result in different trading strategy, including speculating, hedging and passive trade. This research also desire to learn the best impact time of Market Profile, so it designs three kinds of experiments with different investment time period to find the optimal time period. The research result indicates that the experiment predicting three day TAIFEX futures trend has the highest profit on speculating trade. Applying the result to hedging on strangle strategy can effectively reduce the max drawdown. And it will get the higher precision on passive trade which is shorting options. Investors can choose suitable strategies by their profit and risk preference.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070053412
http://hdl.handle.net/11536/71827
顯示於類別:畢業論文