search for


KOSPI200 futures index prediction using denoising filter and LSTM
Journal of the Korean Data & Information Science Society 2019;30:645-54
Published online May 31, 2019;
© 2019 Korean Data and Information Science Society.

Nak Young Lee1 · Kyong Joo Oh2

12Department of Industrial Engineering, Yonsei University
Correspondence to: Professor, Department of Industrial Engineering, Yonsei University, Seoul 03722, Korea. E-mail:
Received April 8, 2019; Revised May 8, 2019; Accepted May 15, 2019.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
There has been many studies which predict the financial market using the deep learning model. However, there has been few studies which apply the denoising filter that improves the performance of predictions removing the noise of financial data. Therefore, the purpose of this study is to apply denoising filter to remove noise from data and then to improve the prediction performance of long short term memory, a deep learning model which is useful for time series prediction. We conducted an empirical analysis using daily and 30 min KOSPI200 futures index data. It is proven that the performance of prediction model using denoising filter is superior to that of the previous long short term memory for the whole period and the sliding window experiment. Also, we confirmed that savitzky-golay filter is more useful for improving the prediction model performance than moving average filter. In the future, denosing filter may be used to improve the prediction performance of various deep learning models.
Keywords : Denoising filter, KOSPI200 future index, LSTM, sliding window.