學術產出-Theses
Article View/Open
Publication Export
-
題名 以序列型學習演算法預測牛市與熊市之轉折點
The sequentially-learning-based algorithm and the pre-diction of the turning points of bull and bear markets作者 張瑄芸
Chang, Hsuan-Yun貢獻者 蔡瑞煌<br>盧敬植
Tsaih, Rua-Huan<br>Lu, Chinh-Chih
張瑄芸
Chang, Hsuan-Yun關鍵詞 概念飄移
離群值偵測
人工神經網路
決策支援機制
移動視窗
concept drifting
outlier detection
artificial neural network
moving window
cramming mechanism
softening mechanism
integrating mechanism日期 2019 上傳時間 7-Aug-2019 16:06:39 (UTC+8) 摘要 在機器學習領域中，對於人工神經網絡（ANN）的架構中，輸入值是實數，輸出值是二元的問題是有挑戰性的，尚未有任何神經網路學習演算法可以解決過度擬合（overfitting）問題，同時可以完美地學習所有訓練數據；另外，在概念飄移環境中要如何去處理離群值的偵測也變得重要，現今的很多資料多為動態性且具概念漂移的特性。為了解決上述挑戰，本研究提出了DSM（決策支持機制）和CSI（強記、軟化、整合）學習演算法。決策支持機制運用了移動視窗的概念，不僅可以識別牛市/熊市中的潛在轉折點檢測，還可以幫助決策者檢視所有轉折點候選者。所提出的CSI學習算法具有以下特點：（1）採用單層隱藏層的神經網絡（ASLFN）和ReLU激活函數；（2）採用LTS原理加速訓練時間；（3）完美的學習所有訓練的數據；（4）實行正則化（regularization），軟化和整合機制，以減輕過度擬合趨勢的模型。我們進行了檢測牛市/熊市轉折點的實驗，以驗證所提出的演算法具有有效性和效率，以及偵測出轉折點候選人幫助決策者作出最終決定。
In the field of machine learning, there is a challenge to the Artificial Neural Networks (ANN) application whose input values are real numbers and the output values are binary. Whether any of ANN learning algorithms can solve the overfitting problem, while it can perfectly learn all of training data. Besides, the problem of outlier detection in the concept environment is becoming an issue. The nature data now has the dynamic and unstable property in the concept drifting environment.To address the aforementioned challenge, this study proposes the DSM (Decision Support Mechanism) and CSI (Cramming, Softening, and Integrating) learning algorithm. DSM apply the moving window mechanism, and it can not only identify the potential turning point detection in the bull/ bear market but also assist the decision maker to double check merely all of turning point candidates. The proposed CSI learning algorithm has the follow-ing features: (1) the adoption of adaptive single-hidden layer feed-forward neural network (ASLFN) and ReLU activation function, (2) the usage of least trimmed squares (LTS) prin-ciple to speed up the training time, (3) the practice to precisely learn all training data, and (4) the implementations of the regularization term, the softening and integrating mechanism to alleviate the obtained model from the overfitting tendency. We conduct an experiment of detecting the turning points of bull/bear markets to validate the effectiveness and efficiency of the proposed algorithm in the addressing challenge.參考文獻 [1] R. R. Trippi, and E. Turban, Neural networks in finance and investing: Using artifi-cial intelligence to improve real world performance, McGraw-Hill, Inc., 1992.[2] H. J. Kim, and K. S. Shin, “A hybrid approach based on neural networks and genetic algorithms for detecting temporal patterns in stock markets,” Applied Soft Computing, vol. 7, no. 2, pp. 569-576, 2007.[3] M. Göçken, M. Özçalıcı, A. Boru, and A. T. Dosdoğru, “Integrating metaheuristics and artificial neural networks for improved stock price prediction,” Expert Systems with Applications, vol. 44, pp. 320-331, 2016.[4] A. Macchiarulo, “Predicting and beating the stock market with machine learning and technical analysis,” Journal of Internet Banking and Commerce, vol. 23, no. 1, pp. 1-22, 2018.[5] E. Chong, C. Han, and F. C. Park, “Deep learning networks for stock market analysis and prediction: Methodology, data representations, and case studies,” Expert Sys-tems with Applications, vol. 83, pp. 187-205, 2017.[6] A. K. Nassirtoussi, S. Aghabozorgi, T. Y. Wah, and D. C. L. Ngo, “Text mining for market prediction: A systematic review,” Expert Systems with Applications, vol. 41, no. 16, pp. 7653-7670, 2014.[7] A. J. Hanna, “A top-down approach to identifying bull and bear market states.,” In-ternational Review of Financial Analysis, vol. 55, pp. 93-110, 2018[8] A. R. Pagan, and K. A. Sossounov, “A simple framework for analysing bull and bear markets,” Journal of Applied Econometrics, vol. 18, no. 1, pp. 23-46, 2003.[9] S. S. Chen, “Predicting the bear stock market: Macroeconomic variables as leading indicators,” Journal of Banking & Finance, vol. 33, no. 2, pp. 211-223, 2009.[10] S. S. Chen, “Revisiting the empirical linkages between stock returns and trading vol-ume,” Journal of Banking & Finance, vol. 36, no. 6, pp. 1781-1788, 2012.[11] A. Tsymbal, “The problem of concept drift: definitions and related work,” Computer Science Department, Trinity College Dublin, vol. 106, no. 2, pp. 58, 2004.[12] R. Elwell, & R. Polikar, “Incremental learning of concept drift in nonstationary envi-ronments,” IEEE Transactions on Neural Networks, vol. 22, no. 10, pp. 1517-1531, 2011.[13] B. Krawczyk, & M. Woźniak, “Diversity measures for one-class classifier ensembles,” Neurocomputing, vol. 126, pp. 36-44, 2014.[14] R. R. Tsaih, “The softening learning procedure,” Mathematical and computer model-ling, vol. 18, no. 8, pp. 61-64, 1993.[15] R. R. Tsaih, “An explanation of reasoning neural networks,” Mathematical and Com-puter Modelling, vol. 28, no. 2, pp. 37-44, 1998.[16] R. H. Tsaih, and T. C. Cheng, “A resistant learning procedure for coping with outli-ers,” Annals of Mathematics and Artificial Intelligence, vol. 57, no. 2, pp. 161-180, 2009.[17] S. Y. Huang, R. H. Tsaih, and F. Yu, “Topological pattern discovery and feature ex-traction for fraudulent financial reporting,” Expert Systems with Applications, vol. 41, no. 9, pp. 4360-4372, 2014.[18] R. H. Tsaih, B. S. Kuo, T. H. Lin, and C. C. Hsu, “The use of big data analytics to predict the foreign exchange rate based on public media: A machine-learning experi-ment,” IT Professional, vol. 20, no. 2, pp. 34-41, 2018.[19] P. J. Rousseeuw, & K. Van Driessen, “Computing LTS regression for large data sets,” Data mining and knowledge discovery, vol. 12, no. 1, pp. 29-45, 2006.[20] D. E. Rumelhart, G. E. Hinton, & J. L. McClelland, “A general framework for parallel distributed processing,” Parallel distributed processing: Explorations in the micro-structure of cognition, vol. 1, no. 45-76, pp. 26, 1986.[21] A. Zaman, P. J. Rousseeuw, & M. Orhan, “Econometric applications of high-breakdown robust regression techniques,” Economics Letters, vol. 71, no. 1, pp. 1-8, 2001.[22] A. C. Atkinson, & T. C. Cheng, “On robust linear regression with incomplete data,” Computational statistics & data analysis, vol. 33, no. 4, pp. 361-380, 2000.[23] J. Gama, I. Žliobaitė, A. Bifet, M. Pechenizkiy, and A. Bouchachia, “A survey on concept drift adaptation,” ACM computing surveys (CSUR), vol. 46, no. 4, pp. 44, 2014.[24] C. W. Lin, A Decision Support Mechanism for Outlier Detection in the Concept Drift-ing Environment, Master`s thesis, 2015.[25] M. Abadi, P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, ... & M. Kudlur, “Tensor-flow: A system for large-scale machine learning,” In 12th {USENIX} Symposium on Operating Systems Design and Implementation, {OSDI} 16, pp. 265-283, 2016.[26] M. Abadi, and TensorFlow, A. A. B. P., “Large-scale machine learning on heterogene-ous distributed systems,” In Proceedings of the 12th USENIX Symposium on Oper-ating Systems Design and Implementation, OSDI’16, Savannah, GA, USA , pp. 265-283, 2016.[27] L. Gonzalez, J. G. Powell, J. Shi, & A. Wilson, “Two centuries of bull and bear mar-ket cycles,” International Review of Economics & Finance, vol. 14 no. 4, pp. 469-486, 2005.[28] G. Bry, & C. Boschan, “Front matter to" Cyclical Analysis of Time Series: Selected Procedures and Computer Programs,” In Cyclical analysis of time series: Selected procedures and computer programs, pp. 13-2. NBER, 1971.[29] B. Babcock, S. Babu, M. Datar, R. Motwani, & J. Widom, “Models and issues in data stream systems,” In Proceedings of the twenty-first ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems, pp. 1-16, ACM, 2002.[30] M. N. Kashani, J. Aminian, S. Shahhosseini,& M. Farrokhi, “Dynamic crude oil foul-ing prediction in industrial preheaters using optimized ANN based moving window technique,” Chemical Engineering Research and Design, vol.90, no.7, pp.938-949, 2012. 描述 碩士
國立政治大學
資訊管理學系
106356017資料來源 http://thesis.lib.nccu.edu.tw/record/#G0106356017 資料類型 thesis dc.contributor.advisor 蔡瑞煌<br>盧敬植 zh_TW dc.contributor.advisor Tsaih, Rua-Huan<br>Lu, Chinh-Chih en_US dc.contributor.author (Authors) 張瑄芸 zh_TW dc.contributor.author (Authors) Chang, Hsuan-Yun en_US dc.creator (作者) 張瑄芸 zh_TW dc.creator (作者) Chang, Hsuan-Yun en_US dc.date (日期) 2019 en_US dc.date.accessioned 7-Aug-2019 16:06:39 (UTC+8) - dc.date.available 7-Aug-2019 16:06:39 (UTC+8) - dc.date.issued (上傳時間) 7-Aug-2019 16:06:39 (UTC+8) - dc.identifier (Other Identifiers) G0106356017 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/124709 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊管理學系 zh_TW dc.description (描述) 106356017 zh_TW dc.description.abstract (摘要) 在機器學習領域中，對於人工神經網絡（ANN）的架構中，輸入值是實數，輸出值是二元的問題是有挑戰性的，尚未有任何神經網路學習演算法可以解決過度擬合（overfitting）問題，同時可以完美地學習所有訓練數據；另外，在概念飄移環境中要如何去處理離群值的偵測也變得重要，現今的很多資料多為動態性且具概念漂移的特性。為了解決上述挑戰，本研究提出了DSM（決策支持機制）和CSI（強記、軟化、整合）學習演算法。決策支持機制運用了移動視窗的概念，不僅可以識別牛市/熊市中的潛在轉折點檢測，還可以幫助決策者檢視所有轉折點候選者。所提出的CSI學習算法具有以下特點：（1）採用單層隱藏層的神經網絡（ASLFN）和ReLU激活函數；（2）採用LTS原理加速訓練時間；（3）完美的學習所有訓練的數據；（4）實行正則化（regularization），軟化和整合機制，以減輕過度擬合趨勢的模型。我們進行了檢測牛市/熊市轉折點的實驗，以驗證所提出的演算法具有有效性和效率，以及偵測出轉折點候選人幫助決策者作出最終決定。 zh_TW dc.description.abstract (摘要) In the field of machine learning, there is a challenge to the Artificial Neural Networks (ANN) application whose input values are real numbers and the output values are binary. Whether any of ANN learning algorithms can solve the overfitting problem, while it can perfectly learn all of training data. Besides, the problem of outlier detection in the concept environment is becoming an issue. The nature data now has the dynamic and unstable property in the concept drifting environment.To address the aforementioned challenge, this study proposes the DSM (Decision Support Mechanism) and CSI (Cramming, Softening, and Integrating) learning algorithm. DSM apply the moving window mechanism, and it can not only identify the potential turning point detection in the bull/ bear market but also assist the decision maker to double check merely all of turning point candidates. The proposed CSI learning algorithm has the follow-ing features: (1) the adoption of adaptive single-hidden layer feed-forward neural network (ASLFN) and ReLU activation function, (2) the usage of least trimmed squares (LTS) prin-ciple to speed up the training time, (3) the practice to precisely learn all training data, and (4) the implementations of the regularization term, the softening and integrating mechanism to alleviate the obtained model from the overfitting tendency. We conduct an experiment of detecting the turning points of bull/bear markets to validate the effectiveness and efficiency of the proposed algorithm in the addressing challenge. en_US dc.description.tableofcontents 1. Introduction 72. Literature Review 92.1 The prediction of bull and bear markets 92.2 Concept drifting 102.3 The single-hidden layer feed-forward neural networks with single output node …………………………………………………………………………………112.4 The back-propagation learning algorithm associated with SLFN with single output node 122.5 The adaptive single-hidden layer feed-forward neural networks with single output node 132.6 Least Trimmed Squares estimator 142.7 Moving Windows 142.8 TensorFlow & GPU 153. The proposed DSM and CSI learning algorithm 173.1. The proposed DSM 173.2 The proposed CSI learning algorithm 184. Experiment design 254.1. The bull and bear market in US stock market 254.2. Variable description 254.3. The description of collected data 274.4. The proposed learning algorithm and its implementation 295. Experiment results 305.1. The performance evaluation 305.2. The experiment results 306. Conclusion and future work 36Reference 37 zh_TW dc.format.extent 1090335 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0106356017 en_US dc.subject (關鍵詞) 概念飄移 zh_TW dc.subject (關鍵詞) 離群值偵測 zh_TW dc.subject (關鍵詞) 人工神經網路 zh_TW dc.subject (關鍵詞) 決策支援機制 zh_TW dc.subject (關鍵詞) 移動視窗 zh_TW dc.subject (關鍵詞) concept drifting en_US dc.subject (關鍵詞) outlier detection en_US dc.subject (關鍵詞) artificial neural network en_US dc.subject (關鍵詞) moving window en_US dc.subject (關鍵詞) cramming mechanism en_US dc.subject (關鍵詞) softening mechanism en_US dc.subject (關鍵詞) integrating mechanism en_US dc.title (題名) 以序列型學習演算法預測牛市與熊市之轉折點 zh_TW dc.title (題名) The sequentially-learning-based algorithm and the pre-diction of the turning points of bull and bear markets en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] R. R. Trippi, and E. Turban, Neural networks in finance and investing: Using artifi-cial intelligence to improve real world performance, McGraw-Hill, Inc., 1992.[2] H. J. Kim, and K. S. Shin, “A hybrid approach based on neural networks and genetic algorithms for detecting temporal patterns in stock markets,” Applied Soft Computing, vol. 7, no. 2, pp. 569-576, 2007.[3] M. Göçken, M. Özçalıcı, A. Boru, and A. T. Dosdoğru, “Integrating metaheuristics and artificial neural networks for improved stock price prediction,” Expert Systems with Applications, vol. 44, pp. 320-331, 2016.[4] A. Macchiarulo, “Predicting and beating the stock market with machine learning and technical analysis,” Journal of Internet Banking and Commerce, vol. 23, no. 1, pp. 1-22, 2018.[5] E. Chong, C. Han, and F. C. Park, “Deep learning networks for stock market analysis and prediction: Methodology, data representations, and case studies,” Expert Sys-tems with Applications, vol. 83, pp. 187-205, 2017.[6] A. K. Nassirtoussi, S. Aghabozorgi, T. Y. Wah, and D. C. L. Ngo, “Text mining for market prediction: A systematic review,” Expert Systems with Applications, vol. 41, no. 16, pp. 7653-7670, 2014.[7] A. J. Hanna, “A top-down approach to identifying bull and bear market states.,” In-ternational Review of Financial Analysis, vol. 55, pp. 93-110, 2018[8] A. R. Pagan, and K. A. Sossounov, “A simple framework for analysing bull and bear markets,” Journal of Applied Econometrics, vol. 18, no. 1, pp. 23-46, 2003.[9] S. S. Chen, “Predicting the bear stock market: Macroeconomic variables as leading indicators,” Journal of Banking & Finance, vol. 33, no. 2, pp. 211-223, 2009.[10] S. S. Chen, “Revisiting the empirical linkages between stock returns and trading vol-ume,” Journal of Banking & Finance, vol. 36, no. 6, pp. 1781-1788, 2012.[11] A. Tsymbal, “The problem of concept drift: definitions and related work,” Computer Science Department, Trinity College Dublin, vol. 106, no. 2, pp. 58, 2004.[12] R. Elwell, & R. Polikar, “Incremental learning of concept drift in nonstationary envi-ronments,” IEEE Transactions on Neural Networks, vol. 22, no. 10, pp. 1517-1531, 2011.[13] B. Krawczyk, & M. Woźniak, “Diversity measures for one-class classifier ensembles,” Neurocomputing, vol. 126, pp. 36-44, 2014.[14] R. R. Tsaih, “The softening learning procedure,” Mathematical and computer model-ling, vol. 18, no. 8, pp. 61-64, 1993.[15] R. R. Tsaih, “An explanation of reasoning neural networks,” Mathematical and Com-puter Modelling, vol. 28, no. 2, pp. 37-44, 1998.[16] R. H. Tsaih, and T. C. Cheng, “A resistant learning procedure for coping with outli-ers,” Annals of Mathematics and Artificial Intelligence, vol. 57, no. 2, pp. 161-180, 2009.[17] S. Y. Huang, R. H. Tsaih, and F. Yu, “Topological pattern discovery and feature ex-traction for fraudulent financial reporting,” Expert Systems with Applications, vol. 41, no. 9, pp. 4360-4372, 2014.[18] R. H. Tsaih, B. S. Kuo, T. H. Lin, and C. C. Hsu, “The use of big data analytics to predict the foreign exchange rate based on public media: A machine-learning experi-ment,” IT Professional, vol. 20, no. 2, pp. 34-41, 2018.[19] P. J. Rousseeuw, & K. Van Driessen, “Computing LTS regression for large data sets,” Data mining and knowledge discovery, vol. 12, no. 1, pp. 29-45, 2006.[20] D. E. Rumelhart, G. E. Hinton, & J. L. McClelland, “A general framework for parallel distributed processing,” Parallel distributed processing: Explorations in the micro-structure of cognition, vol. 1, no. 45-76, pp. 26, 1986.[21] A. Zaman, P. J. Rousseeuw, & M. Orhan, “Econometric applications of high-breakdown robust regression techniques,” Economics Letters, vol. 71, no. 1, pp. 1-8, 2001.[22] A. C. Atkinson, & T. C. Cheng, “On robust linear regression with incomplete data,” Computational statistics & data analysis, vol. 33, no. 4, pp. 361-380, 2000.[23] J. Gama, I. Žliobaitė, A. Bifet, M. Pechenizkiy, and A. Bouchachia, “A survey on concept drift adaptation,” ACM computing surveys (CSUR), vol. 46, no. 4, pp. 44, 2014.[24] C. W. Lin, A Decision Support Mechanism for Outlier Detection in the Concept Drift-ing Environment, Master`s thesis, 2015.[25] M. Abadi, P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, ... & M. Kudlur, “Tensor-flow: A system for large-scale machine learning,” In 12th {USENIX} Symposium on Operating Systems Design and Implementation, {OSDI} 16, pp. 265-283, 2016.[26] M. Abadi, and TensorFlow, A. A. B. P., “Large-scale machine learning on heterogene-ous distributed systems,” In Proceedings of the 12th USENIX Symposium on Oper-ating Systems Design and Implementation, OSDI’16, Savannah, GA, USA , pp. 265-283, 2016.[27] L. Gonzalez, J. G. Powell, J. Shi, & A. Wilson, “Two centuries of bull and bear mar-ket cycles,” International Review of Economics & Finance, vol. 14 no. 4, pp. 469-486, 2005.[28] G. Bry, & C. Boschan, “Front matter to" Cyclical Analysis of Time Series: Selected Procedures and Computer Programs,” In Cyclical analysis of time series: Selected procedures and computer programs, pp. 13-2. NBER, 1971.[29] B. Babcock, S. Babu, M. Datar, R. Motwani, & J. Widom, “Models and issues in data stream systems,” In Proceedings of the twenty-first ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems, pp. 1-16, ACM, 2002.[30] M. N. Kashani, J. Aminian, S. Shahhosseini,& M. Farrokhi, “Dynamic crude oil foul-ing prediction in industrial preheaters using optimized ANN based moving window technique,” Chemical Engineering Research and Design, vol.90, no.7, pp.938-949, 2012. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU201900325 en_US
