應用技術分析於期貨市場進行危機預警：以台指期為例

Applying Technical Analysis in Futures Market For Crysis Detection: Using Taiwan Stock Price Index Futures As Example

馮竣嗣(Feng,Chun Szu)

指導教授 ： 王俊程 ； 蔡子晧

國立清華大學/科技管理學院/經營管理碩士在職專班/碩士(2016年)

摘要

臺灣證券交易所股價指數期貨於1998年以來成交量呈現百倍的成長，貼近台灣市場以及使用新台幣結算等便利性不容忽視，歷經幾次金融危機震盪起伏，技術分析雖有事後諸葛之嫌，但難以忽視其重要性。
本文以技術分析為主，基因演算法為輔，將市場上的交易情緒忠實呈現，在各個技術分析以及交易指標中取得趨吉避凶的方法，經過實驗及歷史回測均能夠在劇烈的波動中避開危機。
基因演算法在運用中較常見於獲利最大化的探討，本文中的設定以獲利平滑，順勢向上為主，除了能夠危機預警也能夠符合交易員在現實交易中的心理壓力平衡，可見基因演算法與技術分析的搭配使用能夠確實地找出趨勢。

關鍵字

技術分析 ； 基因演算法 ； 獲利平滑 ； 危機預警

並列摘要

Since 1998 , the trading volume of FITX has had hundredfold growth. We should not underestimate the convenience of FITX, for instance, being line with Taiwan market and using New Taiwan dollar to close an account. Although technical analysis after several times of financial turmoil seems like twenty-twenty hindsight, we should not ignore the importance of it.
This thesis mainly rely on technical analysis supplemented by Genetic Algorithm, truly presents investors’ sentiment, finding way to pursue interests and to prevent losses in many kinds of technical analysis and trading indicators. After experiment and backtesting we are also able to avoid crisis in strong volatility.
Genetic Algorithm is commonly used to maximize benefits. In this thesis, Genetic Algorithm is mainly used to get a stable growth. It is able to avoid crisis, and conforming to the investors’ sentiment in real trading. It can be seen that we are really able to find trend by using Genetic Algorithm with technical analysis.

並列關鍵字

Technical analysis ； Genetic Algorithm ； Stable growth ； avoid crisis

參考文獻

1. 林文修、陳仕哲（2014），遺傳演算法在台灣股價趨勢轉折點與 波動訊號捕捉之應用。

5. 徐弘翰（2014），運用基因演算法買賣期貨之研究-以台股期貨為例。

6. 連立川、葉怡成、謝明勳，「以遺傳演算法建構台灣股市買賣決策規則之研究」，2004 智慧型知識經濟研討會暨第二屆演化式計算應用專題研討會，台北市，民國 93 年。

B. Allen, F., and Karjalainen, R., “Using Genetic Algorithms to Find Technical Trading Rules,” Journal of Finance Economics （51）, 1999, pp. 245-271.

C. Armano, Giuliano, et al., “Stock market prediction by a mixture of genetic-neural experts,” International Journal of Pattern Recognition and Artificial Intelligence （16）, 2002, pp.501-526.