學術產出-Theses
Article View/Open
Publication Export
-
題名 應用長短期記憶神經網絡於指數型基金之研究
A Study of ETFs Trading Strategy Using Long Short-Term Memory Neural Networks作者 謝長杰
Hsieh, Chang-Chieh貢獻者 胡毓忠
Hu, Yuh-Jong
謝長杰
Hsieh, Chang-Chieh關鍵詞 交易策略
小波轉換
長短期記憶神經網絡
Trading strategy
Wavelet transform
LSTM日期 2021 上傳時間 4-Aug-2021 16:30:53 (UTC+8) 摘要 近年來，長短期記憶(LSTM)技術被廣泛用於預測金融市場的資產價格走勢。然而，這些研究方法中只有少數可以帶來實際利潤。因此本研究提出了一種新的混合模型，稱為動態WT-FLF-LSTM，它在一定的損失函數下結合了小波變換和LSTM。我們評估了六個主要市場ETF的交易策略。盈利表現在所有市場均有大幅提升。所有市場的最大跌幅都在20%以內，而平均交易日在11到16天之間。這一結果表明我們的模型適用於現實世界的交易。此外，我們的模型在大多數金融市場中的表現優於買入並持有策略的基準。為了顯示我們方法的穩健性，我們在台灣50ETF上測試了長期策略，並獲得了30.82%的年化回報率和1.07的夏普比率。
In recent years, the Long ShortTerm Memory (LSTM) technique widely used to predict asset price movements in the financial market. However, in practice, only a few of these studies’ methods could lead to actual profits. This paper presents a novel hybrid model called dynamic WTFLFLSTM, which combines wavelet transform and LSTM under a certain loss function. We evaluate the trading strategy in six significant markets’ ETF. The profitability performance has a substantial enhancement in all markets. The maximum drawdown in all markets is contained within 20%, while the average trading days are between 11 and 16. This outcome indicates the suitability of our model for real world trading. Furthermore, our model outperforms the benchmark of a buyandhold strategy in most financial markets. To show our method’s robustness, we test the longshot strategy on the Taiwan Top 50 ETF (0050.TW) and obtain an annualized return of 30.82% and a Sharpe ratio 1.07. Our study provides a robust trading system with a lower forecasting error.參考文獻 [1] Ahmed, S., Hassan, S.U., Aljohani, N. R., and Nawaz, R. FLFLSTM: A novel prediction system using Forex loss function. Applied Soft Computing 97, Part B (2020), Article 106780.[2] Alonso-Monsalve, S., Suárez-Cetrulo, A. L., Cervantes, A., and Quintana, D. Convolution on neural networks for high-frequency trend prediction of cryptocurrency exchange rates using technical indicators. Expert Systems with Applications 149 (2020), Article 113250.[3] Bao, W., Yue, J., and Rao, Y. A deep learning framework for financial time series using stacked autoencoders and long-short term memory. PLOS ONE 12, 7 (2017), pp. 1–24.[4] Biazon, V., and Bianchi, R. Gated recurrent unit networks and discrete wavelet transforms applied to forecasting and trading in the stock market. In Anais do XVII Encontro Nacional de Inteligência Artificial e Computacional (2020), pp. 650–661.[5] Borovkova, S., and Tsiamas, I. An ensemble of LSTM neural networks for high-frequency stock market classification. Journal of Forecasting 38, 6 (2019), pp. 600–619.[6] Chen, Y., Wu, J., and Bu, H. Stock market embedding and prediction: A deep learning method. In 2018 15th International Conference on Service Systems and Service Management (ICSSSM) (2018), pp. 1–6.[7] Donoho, D. L., Johnstone, I. M., Kerkyacharian, G., and Picard, D. Wavelet shrinkage: Asymptopia Journal of the Royal Statistical Society: Series B (Methodological) 57, 2 (1995), pp. 301–337.[8] Hiransha, M., Gopalakrishnan, E. A., Menon, V. K., and Soman, K. Nse stock market prediction using deep-learning models. Procedia Computer Science 132 (2018), pp. 1351–1362.[9] Hsieh, T.J., Hsiao, H.F., and Yeh, W.C. Forecasting stock markets using wavelet transforms and recurrent neural networks: An integrated system based on artificial bee colony algorithm. Applied Soft Computing 11, 2 (2011), pp. 2510–2525.[10] Islam, M. S., and Hossain, E. Foreign exchange currency rate prediction using a GRU–LSTM hybrid network. Soft Computing Letters (2020), Article 100009.[11] Li, A. W., and Bastos, G. S. Stock market forecasting using deep learning and technical analysis: A systematic review. IEEE Access 8 (2020), pp. 185232–185242.[12] Li, Z., and Tam, V. Combining the real-time wavelet denoising and long-short-term memory neural network for predicting stock indexes. In 2017 IEEE Symposium Series on Computational Intelligence (2017), pp. 1–8.[13] Livieris, I. E., Pintelas, E., and Pintelas, P. A CNN–LSTM model for gold price timeseries forecasting. Neural Computing and Applications 32, 23 (2020), pp. 17351–17360.[14] Luo, G., and Zhang, D. Recognition of partial discharge using wavelet entropy and neural network for tev measurement. In 2012 IEEE International Conference on Power System Technology (POWERCON) (2012), pp. 1–6.[15] Mallat, S. G. A theory for multiresolution signal decomposition: The wavelet representation. IEEE Transactions on Pattern Analysis and Machine Intelligence 11, 7 (1989), pp. 674–693.[16] Nguyen, T. H., and Shirai, K. Topic modeling based sentiment analysis on social media for stock market prediction. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) (2015), pp. 1354–1364.[17] Pang, X., Zhou, Y., Wang, P., Lin, W., and Chang, V. An innovative neural network approach for stock market prediction. The Journal of Supercomputing 76, 3 (2020), pp.2098–2118.[18] Qiu, J., Wang, B., and Zhou, C. Forecasting stock prices with long-short term memory neural network based on attention mechanism. PLOS ONE 15, 1 (2020), pp. 1–15.[19] Rundo, F. Deep lstm with reinforcement learning layer for financial trend prediction in fx high frequency trading systems. Applied Sciences 9, 20 (2019), 4460.[20] Sengupta, S., Basak, S., Saikia, P., Paul, S., Tsalavoutis, V., Atiah, F., Ravi, V., and Peters, A. A review of deep learning with special emphasis on architectures, applications and recent trends. Knowledge-Based Systems 194 (2020), Article 105596.[21] Sezer, O. B., Gudelek, M. U., and Ozbayoglu, A. M. Financial time series forecasting with deep learning: A systematic literature review: 2005–2019. Applied Soft Computing 90 (2020), Article 106181.[22] Tang, H., Chiu, K.C., and Xu, L. Finite mixture of ARMA-GARCH model for stock price prediction. In Proceedings of the Third International Workshop on Computational Intelligence in Economics and Finance (CIEF’2003), North Carolina, USA (2003), pp. 1112–1119.[23] Wong, W., Xu, L., and Yip, F. Financial prediction by finite mixture GARCH model. In Proceedings of Fifth International Conference on Neural Information Processing: ICONIP 98, Kitakyushu, Japan (1998), pp. 1351–1354. 描述 碩士
國立政治大學
資訊科學系碩士在職專班
106971014資料來源 http://thesis.lib.nccu.edu.tw/record/#G0106971014 資料類型 thesis dc.contributor.advisor 胡毓忠 zh_TW dc.contributor.advisor Hu, Yuh-Jong en_US dc.contributor.author (Authors) 謝長杰 zh_TW dc.contributor.author (Authors) Hsieh, Chang-Chieh en_US dc.creator (作者) 謝長杰 zh_TW dc.creator (作者) Hsieh, Chang-Chieh en_US dc.date (日期) 2021 en_US dc.date.accessioned 4-Aug-2021 16:30:53 (UTC+8) - dc.date.available 4-Aug-2021 16:30:53 (UTC+8) - dc.date.issued (上傳時間) 4-Aug-2021 16:30:53 (UTC+8) - dc.identifier (Other Identifiers) G0106971014 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/136704 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系碩士在職專班 zh_TW dc.description (描述) 106971014 zh_TW dc.description.abstract (摘要) 近年來，長短期記憶(LSTM)技術被廣泛用於預測金融市場的資產價格走勢。然而，這些研究方法中只有少數可以帶來實際利潤。因此本研究提出了一種新的混合模型，稱為動態WT-FLF-LSTM，它在一定的損失函數下結合了小波變換和LSTM。我們評估了六個主要市場ETF的交易策略。盈利表現在所有市場均有大幅提升。所有市場的最大跌幅都在20%以內，而平均交易日在11到16天之間。這一結果表明我們的模型適用於現實世界的交易。此外，我們的模型在大多數金融市場中的表現優於買入並持有策略的基準。為了顯示我們方法的穩健性，我們在台灣50ETF上測試了長期策略，並獲得了30.82%的年化回報率和1.07的夏普比率。 zh_TW dc.description.abstract (摘要) In recent years, the Long ShortTerm Memory (LSTM) technique widely used to predict asset price movements in the financial market. However, in practice, only a few of these studies’ methods could lead to actual profits. This paper presents a novel hybrid model called dynamic WTFLFLSTM, which combines wavelet transform and LSTM under a certain loss function. We evaluate the trading strategy in six significant markets’ ETF. The profitability performance has a substantial enhancement in all markets. The maximum drawdown in all markets is contained within 20%, while the average trading days are between 11 and 16. This outcome indicates the suitability of our model for real world trading. Furthermore, our model outperforms the benchmark of a buyandhold strategy in most financial markets. To show our method’s robustness, we test the longshot strategy on the Taiwan Top 50 ETF (0050.TW) and obtain an annualized return of 30.82% and a Sharpe ratio 1.07. Our study provides a robust trading system with a lower forecasting error. en_US dc.description.tableofcontents 1 Introduction 12 Literature 33 Methodology 63.1 Wavelet Transform 63.2 Long Short-Term Memory (LSTM) 93.2.1 LSTM Algorithm 93.2.2 Forex Loss Function (FLF) 113.3 Trading Strategy 123.4 System Pipeline 144 Empirical Analysis 174.1 Data Description 174.2 Profitability Comparison 184.3 Mean Square Error Comparison 274.4 Profitability of the Long-short Strategy 275 Conclusions 336 Future Work 35Reference 36 zh_TW dc.format.extent 1584877 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0106971014 en_US dc.subject (關鍵詞) 交易策略 zh_TW dc.subject (關鍵詞) 小波轉換 zh_TW dc.subject (關鍵詞) 長短期記憶神經網絡 zh_TW dc.subject (關鍵詞) Trading strategy en_US dc.subject (關鍵詞) Wavelet transform en_US dc.subject (關鍵詞) LSTM en_US dc.title (題名) 應用長短期記憶神經網絡於指數型基金之研究 zh_TW dc.title (題名) A Study of ETFs Trading Strategy Using Long Short-Term Memory Neural Networks en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] Ahmed, S., Hassan, S.U., Aljohani, N. R., and Nawaz, R. FLFLSTM: A novel prediction system using Forex loss function. Applied Soft Computing 97, Part B (2020), Article 106780.[2] Alonso-Monsalve, S., Suárez-Cetrulo, A. L., Cervantes, A., and Quintana, D. Convolution on neural networks for high-frequency trend prediction of cryptocurrency exchange rates using technical indicators. Expert Systems with Applications 149 (2020), Article 113250.[3] Bao, W., Yue, J., and Rao, Y. A deep learning framework for financial time series using stacked autoencoders and long-short term memory. PLOS ONE 12, 7 (2017), pp. 1–24.[4] Biazon, V., and Bianchi, R. Gated recurrent unit networks and discrete wavelet transforms applied to forecasting and trading in the stock market. In Anais do XVII Encontro Nacional de Inteligência Artificial e Computacional (2020), pp. 650–661.[5] Borovkova, S., and Tsiamas, I. An ensemble of LSTM neural networks for high-frequency stock market classification. Journal of Forecasting 38, 6 (2019), pp. 600–619.[6] Chen, Y., Wu, J., and Bu, H. Stock market embedding and prediction: A deep learning method. In 2018 15th International Conference on Service Systems and Service Management (ICSSSM) (2018), pp. 1–6.[7] Donoho, D. L., Johnstone, I. M., Kerkyacharian, G., and Picard, D. Wavelet shrinkage: Asymptopia Journal of the Royal Statistical Society: Series B (Methodological) 57, 2 (1995), pp. 301–337.[8] Hiransha, M., Gopalakrishnan, E. A., Menon, V. K., and Soman, K. Nse stock market prediction using deep-learning models. Procedia Computer Science 132 (2018), pp. 1351–1362.[9] Hsieh, T.J., Hsiao, H.F., and Yeh, W.C. Forecasting stock markets using wavelet transforms and recurrent neural networks: An integrated system based on artificial bee colony algorithm. Applied Soft Computing 11, 2 (2011), pp. 2510–2525.[10] Islam, M. S., and Hossain, E. Foreign exchange currency rate prediction using a GRU–LSTM hybrid network. Soft Computing Letters (2020), Article 100009.[11] Li, A. W., and Bastos, G. S. Stock market forecasting using deep learning and technical analysis: A systematic review. IEEE Access 8 (2020), pp. 185232–185242.[12] Li, Z., and Tam, V. Combining the real-time wavelet denoising and long-short-term memory neural network for predicting stock indexes. In 2017 IEEE Symposium Series on Computational Intelligence (2017), pp. 1–8.[13] Livieris, I. E., Pintelas, E., and Pintelas, P. A CNN–LSTM model for gold price timeseries forecasting. Neural Computing and Applications 32, 23 (2020), pp. 17351–17360.[14] Luo, G., and Zhang, D. Recognition of partial discharge using wavelet entropy and neural network for tev measurement. In 2012 IEEE International Conference on Power System Technology (POWERCON) (2012), pp. 1–6.[15] Mallat, S. G. A theory for multiresolution signal decomposition: The wavelet representation. IEEE Transactions on Pattern Analysis and Machine Intelligence 11, 7 (1989), pp. 674–693.[16] Nguyen, T. H., and Shirai, K. Topic modeling based sentiment analysis on social media for stock market prediction. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) (2015), pp. 1354–1364.[17] Pang, X., Zhou, Y., Wang, P., Lin, W., and Chang, V. An innovative neural network approach for stock market prediction. The Journal of Supercomputing 76, 3 (2020), pp.2098–2118.[18] Qiu, J., Wang, B., and Zhou, C. Forecasting stock prices with long-short term memory neural network based on attention mechanism. PLOS ONE 15, 1 (2020), pp. 1–15.[19] Rundo, F. Deep lstm with reinforcement learning layer for financial trend prediction in fx high frequency trading systems. Applied Sciences 9, 20 (2019), 4460.[20] Sengupta, S., Basak, S., Saikia, P., Paul, S., Tsalavoutis, V., Atiah, F., Ravi, V., and Peters, A. A review of deep learning with special emphasis on architectures, applications and recent trends. Knowledge-Based Systems 194 (2020), Article 105596.[21] Sezer, O. B., Gudelek, M. U., and Ozbayoglu, A. M. Financial time series forecasting with deep learning: A systematic literature review: 2005–2019. Applied Soft Computing 90 (2020), Article 106181.[22] Tang, H., Chiu, K.C., and Xu, L. Finite mixture of ARMA-GARCH model for stock price prediction. In Proceedings of the Third International Workshop on Computational Intelligence in Economics and Finance (CIEF’2003), North Carolina, USA (2003), pp. 1112–1119.[23] Wong, W., Xu, L., and Yip, F. Financial prediction by finite mixture GARCH model. In Proceedings of Fifth International Conference on Neural Information Processing: ICONIP 98, Kitakyushu, Japan (1998), pp. 1351–1354. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202100813 en_US
