結合牛式樣板與技術指標預測臺灣股市指數之漲跌

本論文研製之牛式樣板與技術指標預測臺股指數之漲跌。本論文依據William Leigh(2002)等人所建構出來之10x10牛式樣板矩陣以及混合各種技術指標，在經過類神經演算法之cross-validation計算後所輸出之結果比較。本論文之實驗樣本為1991/01/01-2012/07/31 之臺灣股市加權指數，資料來源取得為臺灣新經濟報(TEJ+)資料庫，本研究發現，在混合指標代入類神經網路之輸出效果最佳，且在9/1000之門檻值設定下，有72.9%之預測準確度。本論文就股市技術面觀點出發，經由實證結果發現，在扣除掉交易成本後，仍可獲得超額利潤。

關鍵字

牛式樣板；類神經網路；技術指標

並列摘要

This paper forecasts TAIEX composite index with "bull flag" and technical indicators. TAIEX composite index data between 1991Q1 to 2012Q2 is retrieved from TEJ+ database, and preprocess with the 10x10 bull flag template matrix approach, technical analysis and a hybrid method. The neural network algorithm cross-validation of the results calculated by comparing the output, result in suggestions on whether transactions on each day would be profitable. With current TAIEX daily price movement limit (0.9%), the hybrid method reached 72.9% accuracy on daily profitability, which is better than template matrix and technical analysis methods. The experiment result shows that the hybrid method could bring excess profit after deducting transaction cost (0.5%).

並列關鍵字

Bull Flag ； Neural Network ； Technical Indicators

參考文獻

1. William Leigh , Russell Purvis , James M. Ragusa “Forecasting the NYSE composite index with technical analysis,pattern recognizer, neural network, and genetic algorithm:a case study in romantic decision support” Decision Support Systems 32 (2002) 361– 377

2. Lawrence Blume, David Easley, Maureen O'Hara, Market Statistics and Technical Analysis: The Role of Volume, Journal of Finance, Volume 49(1994), 153-181

4. 陳執中，臺股加權指數隔月收盤價預測之研究，國立成功大學統計學研究所碩士論文，2006

10. 李惠妍、吳宗正、溫敏杰，迴歸模式與類神經網路在臺股指數期貨預測之研究，經營管理論叢 Vol.2, No.1 2006，第83-99頁，2006

3. Ron Kohavi, A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection, Appears in the International Joint Conference on Artificial Intelligence(IJCAI) .1995

Google Scholar

國際替代計量

結合牛式樣板與技術指標預測臺灣股市指數之漲跌

全文下載

主題瀏覽