It's been more than 90 days since you changed your password. Please change your password to protect your privacy
When you visit DBpia through the library, login to SNS is restricted.
If you visit through the library, the URL will be different and you will not be able to log in to SNS due to kakao/google/naver policy. If you want to use your SNS account, please visit DBpia directly and log in.
This study proposed a deep learning parallel model for stock price prediction using stock price data, technical analysis data, and environmental factor data. The data set for prediction was divided into three, data set 1 is the opening price, high price, low price, closing price, and trading volume, data set 2 added technical analysis data, and data set 3 is the exchange rate that can affect the stock price. the overall industrial production index was added. The deep learning model proposed DNN, LSTM, and 1D-CNN models as basic models, and a DCNN model that merged 1D-CNN based on the DNN model as a parallel model, and a DLSTM model that merged LSTM as a parallel model. As a result of the experiment, the performance of LSTM and BiLSTM models was higher than that of DNN and CNN, and in particular, the DLSTM model, a parallel model, performed the best. For the DLSTM model, which is a parallel model, the RMSE of data set 1 was 0.0091, the RMSE of data set 2 was 0.0080, and the RMSE of data set 3 was 0.0071. Data set 3, which combined all data, had the best performance.
본 연구는 주가 데이터, 기술 분석 데이터, 환경요소 데이터를 이용하여, 주가예측을 위한 딥러닝 병렬 모델을 제안하였다. 예측을 위한 데이터 셋은 3개로 나누었으며, 데이터 셋 1은 시가, 고가, 저가, 종가, 거래량이며, 데이터 셋 2는 기술 분석 데이터를 추가하였으며, 데이터 셋 3은 주가에 영향을 줄 수 있는 환율, 전산업 생산지수를 추가하였다. 딥러닝 모델은 기본 모델로서 DNN, LSTM, 1D-CNN 모델과 병렬 모델로서 DNN 모델을 기본으로 1D-CNN을 병합한 DCNN 모델과 LSTM을 병합한 DLSTM 모델을 제안하였다. 실험 결과, DNN과 CNN 보다는 LSTM과 BiLSTM 모델의 성능이 높았으며, 특히 병렬모델인 DLSTM 모델이 가장 성능이 좋았다. 병렬 모델인 DLSTM 모델에 대한 데이터 셋 1의 RMSE는 0.0091, 데이터 셋 2의 RMSE는 0.0080, 데이터 셋 3의 RMSE는 0.0071로서 모든 데이터가 합쳐진 데이터 셋 3의 성능이 가장 좋았다.
You have been reported one after another from other members for writing reviews that contain advertisements, busive language, personal attacks, or disparaging expressions. As a result, you cannot write a review.