根據此篇Paper: Genetic Algorithms for the Investment of the Mutual Fund with Global Trend Indicator為出發點，試圖更貼近基金買賣時的決策與行為模式，主要提出了四點改善去修改此篇paper: (1)修改GTI的計算方式，避免原先全上跌與全上漲，可能造成的問題，(2)加入容忍度，每次都是挑前面名次去投資限制下，介於門檻附近的基金，可能會過於頻繁交換名次，交易費間接造成獲利減少，(3)加入Stop-Loss point觀念，在Sell時，動態去釋放手上持有基金，來代替原本賣空，(4)有人喜好看短期數據做投資以求更高獲利，但相對提高風險，也有喜好看長期，所以加入(1-α)History + (α)Recent，讓使用者自行設定，並設計基因演算法去求得α以供參考。

在最後的實驗結果，與原論文中，以2007年最好的投資報酬率(ROI)實驗結果8.98%做相互比較，在我設立三種不同停損與釋放的係數下，結果都比原方法高出四倍的投資報酬率。

This research proposes a decision and behavior model which tries to approximate the fund trading. The main idea is based on the principle of the publication “Genetic Algorithms for the Investment of the Mutual Fund with Global Trend Indicator”, and four optimization schemes are proposed as well. First, the calculation of GTI is refined to prevent the possible problems caused by the case that all the fund are getting rise, or the opposite. Second, the tolerance is considered to avoid the reduction of profits owing to the increase of rates for transaction which Funds, those near threshold ones, might exchange ranking too often. Third, the concept of Stop-Loss Point is involved to release the fund dynamically instead of oversell. The last, Someone like to investment more profitable with short-term data, but high-risk. Someone like to investment long-term data, therefore, we added (1-α)History + (α)Recent to make users could set by themselves. And we also design genetic algorithm to calculate α for reference.

Under the constraints of three different coefficients of stop-loss and release, the Return of Investment (ROI) is four times than original one(8.98%), which is compared in 2007.

序言 _______________________________________________________________
致謝....................................................................................................................................i
中文摘要...........................................................................................................................ii
英文摘要..........................................................................................................................iii
目錄..................................................................................................................................iv

正文 _______________________________________________________________
一. 序論...........................................................................................................................1
1.1. 動機與目的.......................................................................................................1
1.2. 大綱與架構.......................................................................................................2
二. Related works............................................................................................................3
2.1. Genetic Algorithm.............................................................................................3
2.1.1.主要流程..................……………………………………………………3
2.1.2. Generation method...…………………………………………………...5
2.2. Related Paper.....................................................................................................7
2.2.1.主要流程..................................................................................................7
2.2.2. Bit string表示物體................................................................................10
2.2.3. Global Trend Indicator(GTI)………………………………………….10
2.2.4. Financial situation……………………………………………………..12
2.2.5. Fitness Function…………………………………………………….…13
2.2.6.實驗方式…...………………………………………………………….14
三. 提出方法.................................................................................................................16
3.1. Improvement GTI...........................................………………………….........16
3.2. 買賣策略........................................................................……………….........19
3.3. (1-α)History + (α)Recent…………………………………………………….27
3.3.1 .Recent…………………………………………………………………28
3.3.2 .History和Recent合併.........................................................................28
四. 實驗.........................................................................................................................30
4.1. 實驗設定............................................. …………………………………..….30
4.2. 計算ROI方式........................................... ……………………………...….31
4.2.1.實驗設計................................................................................................31
4.3. 實驗結果......................................... ………………………………………...32
五. 結論....................................................... ………………………………………….34
六. Future works................................................... …………………………………....36
參考資料. ……………………………………………...................................................38

[1]http://zh.wikipedia.org/wiki/%E9%81%BA%E5%82%B3%E7%AE%97%E6%B3%95
[2] Johann Dreo, “Metaheuristics for hard optimization”, 2006
[3] J. H. Holland,” Adaptation in Natural and Artificial Systems”, 1975.
[4] H. H. Chen, C. B. Yang, and Y. H. Peng, “Genetic programming for the investment of the mutual fund with sortino ratio and mean variance model,” Proc. of the 15th Conference on Artificial Intelligence and Applications, Hsinchu, Taiwan, Nov. 2010.
[5] T. J. Tsai, C. B. Yang, and Y. H. Peng, “Genetic algorithms for the investment of the mutual fund with global trend indicator,” Expert Systems with Applications, Vol. 38(3), pp. 1697–1701, 2011
[6] P. Y. Chen, C. H. Lin, “An intelligent system for predicting stock trading strategies using case-based reasoning and neural network”, 2009
[7] 蕭碧燕, “買基金為自己加薪”, 2007
[8] 楊仁和 譯, “深入淺出統計學”, 2009
[9] 邱顯比, “基金理財的六堂課”,2002
[10] 邱顯比, 朱成志, “少犯錯, 一生都是投資贏家”, 2010
[11] John J. Murphy, “Technical Analysis of the Financial Markets”, 2000
[12] 張聖傑, “買錯基金，也賺錢？！”, 2005
[13] Fund DJ, http://www.funddj.com/, 2000.
[14] 基智網, http://www.moneydj.com/funddj/
[15] 楊仁和 譯, “深入淺出統計學”, 2009
[16]http://zh.wikipedia.org/wiki/%E5%85%B1%E5%90%8C%E5%9F%BA%E9%87%91