台灣指數期貨跳空開盤後價格反轉行為研究

Abstract

期貨市場和其他商品市場是一樣由買賣雙方競價交易的結果，買方和賣方在某一價格下達成交易。而參與期貨市場買賣的也是人所組成，當期貨價格下跌時會吸引更多買方進場購買，當期貨價格上漲時則會吸引更多賣方進場。 然而，不管是股票市場或期貨市場都是瞬息萬變，市場消息千變萬化，價格快速波動。投資人通常無法掌握正確市場價格走向而錯過進場時機，常常造成追高及追低的情況，造成投資人巨大損失。 本研究嘗試透過期貨市場的開盤價、最高價、最低價、收盤價所產生的K線及影線、輔以市場輪廓理論之觀念，設計一種能夠動態偵測期貨市場發生跳空開盤後市場價格趨勢反轉進場點的模型。 最後，再結合成交量的相關資訊作為倒傳遞類神經網路的輸入值，期望利用倒傳遞類神經網路的學習能夠進一步分析此模型的精確性。 由實驗結果顯示，以成交量及多空雙方物理量為倒傳遞類神經網路的輸入參數時，能夠更準確分析出市場在跳空開盤後價格出現反轉趨勢。且在短線持有的情況下準確率相對更高。

Futures market and other commodity markets both are a result of bidding price transactions between buyers and sellers and they conclude a transaction. And then in the futures market trading is composed of buyers and sellers. When the price of futures is falling it will attract more buyers enter the market for buying, and on the other hand when the futures are rising it will attract more sellers enter the market. However, the price fluctuation varies from minute to minute according to the stock market or futures market or the market news. Normally, it is hard for investors to grasp the tendency of correct market price, they may chase the lowest or highest price but that cause the huge losses of investors. This research attempt through the open, high, low and close price of futures market with the result from candlestick chart, shadow, and market profile to design a model that it can dynamically detect the trend reversal after gapped open in futures market. Eventually, the model combine with the information of volume can be made as a value of back-propagation neural network. And except after use back-propagation neural network as research that we can analysis more accuracy of the model. From the experiment result shows the volume with both long and short physics property can be made as a value of back-propagation neural network. And it can analysis more accuracy of trend reversal after gapped open in futures market, especial in the case of short-term trading will be more accuracy.