(3.236.175.108) 您好!臺灣時間:2021/03/01 12:31
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:施榮恩
研究生(外文):SHIH,RONG-EN
論文名稱:機器學習建構之交易策略與未來失效風險關係研究-以Strategy Quant 工具開發之策略為例
論文名稱(外文):Relationship Between Trading Strategies Built by Machine learning with Future FailureRisk. An Example of Strategies Built by Strategy Quant
指導教授:洪志興洪志興引用關係
指導教授(外文):HUNG, CHIH-HSING
口試委員:臧仕維李慶章洪志興王銘駿陳勤明
口試委員(外文):TZANG, SHYH-WEIRLEE, CHING-CHANGHUNG, CHIH-HSINGWANG, MING-CHUNCHEN, CHIN-MING
口試日期:2020-07-12
學位類別:碩士
校院名稱:國立高雄科技大學
系所名稱:金融系
學門:商業及管理學門
學類:財務金融學類
論文種類:學術論文
論文出版年:2020
畢業學年度:108
語文別:中文
論文頁數:49
中文關鍵詞:機器學習交易量化交易技術分析時間序列分析
外文關鍵詞:Machine learning tradingalgorithmic tradingtechnical analysistime serials analysis
相關次數:
  • 被引用被引用:0
  • 點閱點閱:115
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
以往股市、期貨、外匯操盤交易,多以人工操作依照基本面或技術面指標、
線型及籌碼面判斷,但常受限於正確性、一致性或可檢驗之邏輯性。在自動交易
發明之後,往往將人工交易策略規則編寫成交易程式,再以過往歷史資料進行回
測,其自動交易程式除多數在歷史回測失效外,編寫有效成功之交易程式與策略
與過程往往曠日費時,當有效之交易程式在過往歷史資料回測成功後,大多數亦
無法通過未來行情之劇變。本研究所使用Strategy Quant 交易機器學習工具以機
器學習方式取代過往人工主觀交易及程式交易,使用主段歷史價格資料做為樣本
內(In sample Training)來訓練有效之交易策略機器人並於次段歷史價格資料做為樣
本外(Out of sample )來觀察模擬這些策略機器人群之交易是否適用於未來或未知
環境。研究結果顯示樣本內訓練期位置與未來失效風險有高度相關,推進分析舉
陣法可預測未來策略存活機率及策略篩選,部份量化指標有助於策略篩選及預測
未來存活率。
Historically, financial instruments, such as stocks, futures, and forex, were traded using fundamental analysis. The methods were incomplete and were supplemented by technical indicators, an innovation favored in today’s financial markets. Trading built on fundamental and technical analysis still lacks verifiability, usefulness, consistency, stability, and discipline. Traditional analysis can avoid human mistakes and poor discipline. But popular trading strategies depend on time-consuming manual rules. This study examines machine-learning tools, or quant strategies, and compares them to manual
trading. The methodology uses in sample training and out of sample back testing. The goal was to observe survivable strategies for unknown trading environments. The results show that the present Sample position in training has high significance with future survival rates, Walk-forward Matrix, and some algorithmic scales, which are able to predict the survival of future and filter strategies.
CONTENTS
摘 要 ............................................................................................................................................ i
Abstract ........................................................................................................................................ ii
Acknowledgments ...................................................................................................................... iii
CONTENTS ................................................................................................................................. iv
LIST OF FIGURES ...................................................................................................................... vi
LIST OF TABLES ...................................................................................................................... vii
Introduction ............................................................................................................. 1
Motivation .......................................................................................................................... 1
Purpose ............................................................................................................................... 2
Research Framework ......................................................................................................... 3
Literature Review .................................................................................................... 5
Program trading ................................................................................................................. 5
Machine Learning .............................................................................................................. 6
Recent Studies .................................................................................................................... 6
Research Methods ................................................................................................... 8
Tools and data .................................................................................................................... 8
Grouping ............................................................................................................................ 9
Strategies ranking and building ....................................................................................... 10
Strategies for In-Sample period Training ......................................................................... 12
Strategies Back Testing in Out of Sample ....................................................................... 17
Robustness tests ............................................................................................................... 17
Predicting in Future period .............................................................................................. 18
Research Results .................................................................................................... 19
Predicting Survival Situations .......................................................................................... 19
v
Monte Carlo testing result with future predicting ............................................................ 25
Walk forward Matrix testing result with future predicting .............................................. 27
Algorithmic scale with future predicting ......................................................................... 28
Conclusions and Suggestions ................................................................................ 35
Strategy life cycle ............................................................................................................ 35
Training Sample Positions ............................................................................................... 35
Strategy filtering and predicting ...................................................................................... 36
References ................................................................................................................................... 38
Froiegn References
[1] Alkhatib, K., Najadat, H., Hmeidi, I., Shatnawi, M. K. A., 2013, "Stock price
prediction using k-nearest neighbor (kNN) algorithm", International Journal of
Business, Humanities and Technology,Vol. 3, 3, pp. 32-44.
[2] Bailey, D. H., Borwein, J., Salehipour, A., Lopez De Prado, M., Zhu, Q. J., 2015,
"Online tools for demonstration of backtest overfitting", Available at SSRN
2597421.
[3] Finacial Conduct Authority, 2020, Retrieved from https://register.fca.org.uk/
[4] Harvey, C. R., Liu, Y., 2014, "Evaluating trading strategies", The Journal of
Portfolio Management,Vol. 40, 5, pp. 108-118.
[5] Min, J. H., Lee, Y.-C., 2005, "Bankruptcy prediction using support vector machine
with optimal choice of kernel function parameters", Expert systems with
applications,Vol. 28, 4, pp. 603-614.
[6] Řeha, F., 2016, "Computational Intelligence for Financial Market Prediction".
[7] Strategy Quant, 2020, Retrieved from https://strategyquant.com/
2. Chinese References
[1] 吳彥璋,2019,機器學習與程式交易在建構交易系統之研究-以台灣指數期
貨為例,輔仁大學,碩士論文。
[2] 林俊吉,2005,應用灰聚類進行台股指數期貨程式交易設計,國立臺灣科技
大學,碩士論文。
[3] 林純仁,2011,以高頻交易方法改善成交量加權平均價格演算法交易策略-
系統建構與實證,國立高雄應用科技大學,碩士論文。
[4] 林敬斌,2009,使用增強式學習法改善一個簡易的臺灣股價指數期貨當沖交
易系統,國立臺灣大學,碩士論文。
[5] 姜林杰祐,2009,程式交易: 觀念, 方法, 技術與解決方案, 二 版, 新
陸出版,台北。
[6] 施育霖,2019,以機器學習方法預測股價:以台股金融類股為案例,國立中
興大學,碩士論文。
[7] 張林忠,2014,分析師關鍵報告: 張林忠教你程式交易, 一 版, 寰宇出
39
版,台北。
[8] 莫慶文,李金玲,楊媚帆,2012,運用通道技術指標於台灣指數期貨之實證
研究。
[9] 陳世勳,2018,程式交易期貨投資人學習歷程之研究,東吳大學,碩士論
文。
[10]喻欣凱,2008,運用支援向量機與文字探勘於股價漲跌趨勢之預測,輔仁大
學,碩士論文。
[11]彭康哲,2016,結合程式交易與選股模型的分析平台之設計與實作 – 以美
國股票市場為例,國立中央大學,碩士論文。
[12]黃培琳,2011,程式交易策略、期貨交易和法人交易之實證研究,朝陽科技
大學,碩士論文。
[13]董寶蘭,2010,程式交易策略實證研究-以投資ETF0050 為例,淡江大學,
碩士論文。
[14]黎藍平,2019,應用機器學習於股價預測 :以越 南股票市場為例,國立高
雄科技大學,碩士論文。
[15]蘇恩德,陳俊汎,陳志鴻,2006,“建構模糊理論風控程式交易之實證研
究”,台灣金融財務季刊, 7 卷,1 期,頁 29-53。
電子全文 電子全文(網際網路公開日期:20250717)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔