The Journal of Korean Institute of Information Technology (한국정보기술학회논문지)

Korean Institute of Information Technology (한국정보기술학회)
권호 표지
DOI QR코드

DOI QR Code

Deep Learning Model for Prediction Rate Improvement of Stock Price Using RNN and LSTM

RNN과 LSTM을 이용한 주가 예측율 향상을 위한 딥러닝 모델

  • Received : 2017.08.01
  • Accepted : 2017.09.16
  • Published : 2017.10.31
⟨ Previous Next ⟩

Abstract

Recently, stock price prediction using deep learning has basically used assistance index as a prediction factors. However assistance index is necessary to examine whether it is suitable as prediction factors because it is subjective viewpoint of researcher. In this study, we examine the suitability as prediction factors with various combinations of existing assistance indexes through the R neural network package, and studied the optimal combinations of assistance indexes and environmental prediction factors like exchange rate, exchange rate moving average, and whole industrial production index in order to improve the prediction rate. In addition, we proposed a deep learning model like DNN, RNN, LSTM which have input-output with extracted prediction factors. As a result, most of the assistance indexes decreased the prediction rate and the prediction rate was improved through additional environmental prediction factors. Also, RNN and LSTM, which are time series deep learning networks, were learned quickly and steadily compared to DNN. Although there is a difference by items, the prediction rate improvement is about 15%.

최근 딥러닝을 이용한 주가예측은 기본적으로 보조지표를 예측요소로 사용하고 있으나, 보조지표는 분석가의 주관적인 관점이기 때문에, 예측요소에 대한 적합여부에 대한 검토가 필요하다. 본 연구는 R의 신경망 패키지를 통해, 기존의 보조지표에 대해서 다양한 조합으로 예측요소 적합여부를 검토하고, 예측율 향상을 위해 최적의 보조지표 조합과 환율, 환율 이동평균, 전산업생산지수 등 환경 예측요소들에 대해서 연구하였다. 또한, 추출된 예측요소를 입출력 패턴으로 DNN, RNN, LSTM 등의 딥러닝 모델을 제안하였다. 연구결과 대부분의 보조지표는 예측율을 저하하는 현상이 있었으며, 추가 환경 예측요소를 통해, 예측율이 향상되었다. 또한, DNN에 비해 시계열 딥러닝 네트워크인 RNN과 LSTM이 빠르고 안정적으로 학습하였으며, 종목별로 차이는 있으나 대략 15% 정도의 예측율 향상을 보였다.

Keywords

References

  1. E. J. Lee, C. H. Min, and T. S. Kim, "Development of the KOSPI (Korea Composite Stock Price Index) forecast model using neural network and statistical methods", The Institute of Electronics Engineers of Korea, Computer and Information, Vol 45, No. 5, pp. 95-101, Sep. 2008.
  2. K. Y. Kim and K. R. Lee, "A Study on the Prediction of Stock Price Using Artificial Intelligence System", Korean Journal of Business Administration, Vol. 21, No. 6, pp. 2421-2449, Dec. 2008.
  3. T. Fischer and C. Krauss, "Deep learning with long short-term memory networks for financial market prediction", FAU Discussion Papers in Economics, No. 11, pp. 310-342, May 2017.
  4. J. W. Lee, "A Stock Trading System based on Supervised Learning of Highly Volatile Stock Price Patterns", Journal of Korean Institute of Information Scientists and Engineers, Vol 19, No. 1, pp. 23-29, Jan. 2013.
  5. C. Hsu, "A hybrid procedure for stock price prediction by integrating self-organizing map and genetic programming", Expert Systems with Applications, Vol. 38, No. 11, pp. 14026-14036. Oct. 2011.
  6. Y. K. Kwon, S. S. Choi, and B. R. Moon, "Stock prediction based on financial correlation", Proceedings of the 7th annual conference on Genetic and evolutionary computation. ACM, pp. 2061-2066, Jun. 2005.
  7. S. M. An, "Deep Learning Architectures and Applications", Korea intelligent information system society, intelligent information study, Vol. 22, No. 2, pp. 127-142, Jun. 2016.
  8. M. Y. Woo, S. Y. Park, Y. W. Han, U. C. Park, "Stock Price Estimation Experiments using R Neural Nets", Proceedings of KIIT Summer Conference 2013, pp. 498-500. May 2013
  9. Giles, C. Lee, S. Lawrence, and A. C. Tsoi, "Noisy time series prediction using recurrent neural networks and grammatical inference", Machine learning, Vol. 44, No. 1, pp. 161-183, Jul. 2001. https://doi.org/10.1023/A:1010884214864
  10. K. J. Jeong and J. S. Choi, "Deep Recurrent Neural Networks," Communications of the Korean Institute of Information Scientists and Engineers, Vol. 33, No. 8, pp. 39-43, Aug. 2015.
  11. T. J. Hsieh, H. F. Hsiao, and W. C. Yeh, "Forecasting stock markets using wavelet transforms and recurrent neural networks: An integrated system based on artificial bee colony algorithm", Applied soft computing, Vol. 11. No. 2, pp. 2510-2525, Mar. 2011. https://doi.org/10.1016/j.asoc.2010.09.007
  12. K. H. Park and H. J. Shin, "Stock Price Prediction Based on Time Series Network", Korean Management Science Review, Vol. 28, No. 1, pp. 53-60, Mar. 2011.
  13. S. Jelena, M. Nijole, and M. Algirdas, "High-low Strategy of Portfolio Composition using Evolino RNN Ensembles", Inzinerine Ekonomika-Engineering Economics 2017, Vol. 28, No. 2, pp. 162-169, Apr. 2017.
  14. F. A. Gers, N. N. Schraudolph, and J. Schmidhuber, "Learning precise timing with LSTM recurrent networks", Journal of Machine Learning Research 3, pp. 115-143, Mar. 2002.
  15. http://colah.github.io/posts/2015-08-Understanding-LSTMs/ [Accessed: Jul. 01. 2017]

Cited by

  1. Universal Prediction System Realization Using RNN vol.16, pp.10, 2017, https://doi.org/10.14801/jkiit.2018.16.10.11
  2. Is Deep Learning for Image Recognition Applicable to Stock Market Prediction? vol.2019, pp.None, 2017, https://doi.org/10.1155/2019/4324878
  3. Model Implementation of Reinforcement Learning for Trading Prediction Using Deep Q Network vol.17, pp.4, 2019, https://doi.org/10.14801/jkiit.2019.17.4.1
  4. A Study on Agricultural Price Prediction System based on Deep Learning vol.17, pp.6, 2019, https://doi.org/10.14801/jkiit.2019.17.6.27
  5. Terra MODIS NDVI 및 LST 자료와 RNN-LSTM을 활용한 토양수분 산정 vol.61, pp.6, 2017, https://doi.org/10.5389/ksae.2019.61.6.123
  6. 빅데이터 기반 실시간 불량품 발생 원인 분석 및 설비 교체주기 예측 vol.19, pp.6, 2019, https://doi.org/10.7236/jiibc.2019.19.6.203
  7. NARX 신경망 최적화를 통한 주가 예측 및 영향 요인에 관한 연구 vol.21, pp.8, 2017, https://doi.org/10.5762/kais.2020.21.8.572
  8. 미국 금리 스프레드를 이용한 한국 금리 스프레드 예측 모델에 관한 연구 : SVR-앙상블(RNN, LSTM, GRU) 모델 기반 vol.43, pp.3, 2017, https://doi.org/10.11627/jkise.2020.43.3.001
  9. Anomaly Detection in Reservoir Water Level Data Using the LSTM Model Based on Deep Learning vol.21, pp.1, 2017, https://doi.org/10.9798/kosham.2021.21.1.71
  10. Catboost 알고리즘을 통한 교통흐름 예측에 관한 연구 vol.22, pp.3, 2021, https://doi.org/10.5762/kais.2021.22.3.58
  11. 딥러닝을 이용한 비트코인 투자전략의 성과 분석 vol.12, pp.4, 2021, https://doi.org/10.15207/jkcs.2021.12.4.249