DOI QR코드

DOI QR Code

A Performance Analysis by Adjusting Learning Methods in Stock Price Prediction Model Using LSTM

LSTM을 이용한 주가예측 모델의 학습방법에 따른 성능분석

  • Jung, Jongjin (Division of Human IT Convergence, Daejin University) ;
  • Kim, Jiyeon (College of Humanities and Arts, Daejin University)
  • 정종진 (대진대학교 휴먼IT융합학부) ;
  • 김지연 (대진대학교 창의미래인재대학)
  • Received : 2020.08.25
  • Accepted : 2020.11.20
  • Published : 2020.11.28

Abstract

Many developments have been steadily carried out by researchers with applying knowledge-based expert system or machine learning algorithms to the financial field. In particular, it is now common to perform knowledge based system trading in using stock prices. Recently, deep learning technologies have been applied to real fields of stock trading marketplace as GPU performance and large scaled data have been supported enough. Especially, LSTM has been tried to apply to stock price prediction because of its compatibility for time series data. In this paper, we implement stock price prediction using LSTM. In modeling of LSTM, we propose a fitness combination of model parameters and activation functions for best performance. Specifically, we propose suitable selection methods of initializers of weights and bias, regularizers to avoid over-fitting, activation functions and optimization methods. We also compare model performances according to the different selections of the above important modeling considering factors on the real-world stock price data of global major companies. Finally, our experimental work brings a fitness method of applying LSTM model to stock price prediction.

과거 인공지능 분야에서는 지식 기반의 전문가 시스템 및 머신러닝 알고리즘들을 금융 분야에 적용하는 연구가 꾸준하게 수행되어 왔다. 특히 주식에 대한 지식 기반의 시스템 트레이딩은 이제 보편화되었고, 최근에는 대용량 데이터에 기반한 딥러닝 기술을 주가 예측에 적용하기 시작했다. 이중 LSTM은 시계열 데이터에 대한 검증된 모델로서 주가 예측에도 적용되고 있다. 본 논문에서는 주가 예측 모델로서 LSTM을 적용할 때 성능향상을 위해 고려해야 할 복잡한 매개변수 설정과 적용 함수들에 대해 적합한 조합 방법을 제안하도록 한다. 크게 가중치와 바이어스에 대한 초기화 대상과 설정 방법, 과적합을 피하기 위한 정규화 적용 대상과 설정 방법, 활성화 함수 적용 방법, 최적화 알고리즘 선택 등을 제시한다. 이 때 나스닥 상장사들에 대한 대용량 데이터를 바탕으로 각각의 방법들을 적용하여 정확도를 비교하면서 평가한다. 이를 통해 주가 예측을 위한 LSTM 적용 시 최적의 모델링 방법을 실증적인 형태로 제안하여 현실적인 시사점을 갖도록 한다. 향후에는 입력 데이터의 포맷과 길이, 하이퍼파라미터들에 대한 성능평가를 추가 수행하여 주요 설정 항목들의 조합에 대한 일반화 연구를 수행하고자 한다.

Keywords

References

  1. Alizadeh, M., et al. (2011). An adaptive neuro fuzzy system for portfolio analysis, International Journal of Intelligent Systems, 22(2), 99-114. https://doi.org/10.1002/int.20456
  2. Behnoush Shakeri et al (2015). Fuzzy Clustering Rule-Based Expert System for Stock Price Movement Prediction, NAFIPS, Redmond, Washington, USA, August 17-19.
  3. R. Lakshman Naik & D. Ramesh & B. Manjula & Dr. A. Govardhan (2012). Prediction of Stock Market Index Using Genetic Algorithm, Computer Engineering and Intelligent Systems, 3(7).
  4. H. J. Kim et al. (2018). Stock Price Prediction Using Deep Learning Ensemble, SIGDB 34(2), 113-120.
  5. S. W. Kim & H. C Ahn (2010). Development of an Intelligent Trading System Using Support Vector Machines and Genetic Algorithms, Journal of Intelligence and Information Systems, 16(1), 71-92.
  6. J. Y. Heo & J. Y. Yang. (2015). SVM based Stock Price Forecasting Using Financial Statements, KIISE transactons on computing practices, 21(3), 167-172. https://doi.org/10.5626/KTCP.2015.21.3.167
  7. G. B. Nam et al. (2017). Development of Stock Investment System Using Machine Learning, 2017 Proceeding of Information Processing Society Fall Conference 24(2), 810-812.
  8. Mao, H. Zeng & X. J, Leng, & G, Zhai (2011). Twitter mood predicts the stock market, Journal of Computational Science, 2(1), 1-8. https://doi.org/10.1016/j.jocs.2010.12.007
  9. Y. J. Song & J. W. Lee & J. W. Lee (2017). Performance Evaluation of Price-based Input Featuress in Stock Price Prediction using Tensorflow, KIISE transactons on computing practices, 23(11).
  10. D. H. Shin, K. H. Choi & C. B. Kim. (2017). Deep Learning Model for Prediction Rate Improvement of Stock Price Using RNN and LSTM, Jounral of KIIT, 15(10), 9-16.
  11. Ashwin Siripurapu (2015). Convolutional Networks for stock Trading, Stanford University.
  12. H. J. Kim et al. (2018). Stock Price Prediction Using Deep Learning Ensemble, SIGDB 34(2), 113-120.
  13. W. S. Lee. (2017). A Deep Learning Analysis of the KOSPI's directions, Jounral of the Korean Data & Information Science Society, 28(2), 287-295. https://doi.org/10.7465/jkdi.2017.28.2.287
  14. T. W. Kim & H. Y. Kim (2019). Forecasting stock prices with a feature fusion LSTM-CNN model using different representations of the same data, PLoS ONE, 14(2).
  15. Klaus Greff et al. (2017). LSTM: A Search Space Odyssey, IEEE Transactions on Neural Networks and Learning Systems, 28(10), 2222-2232. https://doi.org/10.1109/TNNLS.2016.2582924
  16. Keras LSTM tutorial - How to easily build a powerful deep learning language model, https://adventuresinmachinelearning/keras-lstmtutorial/
  17. Yahoo Finance. https://finance.yahoo.com6