Yuyang Lin, Qiyin Zhong, Qi Huang, Muyang Li, Fei Ma
{"title":"A new convolutional neural network and long short term memory combined model for stock index prediction","authors":"Yuyang Lin, Qiyin Zhong, Qi Huang, Muyang Li, Fei Ma","doi":"10.1109/CISP-BMEI53629.2021.9624337","DOIUrl":null,"url":null,"abstract":"Stock market is one of the most important parts in the financial market. Numerous time series forecasting methods have been developed for predicting the stock price. Feature extraction is essential to many of these forecasting models. Highly related features can improve the accuracy of the forecasting model. This paper proposes a new model named CNN-LS that combines Convolution Neural Networks (CNN) with Long Short-Term Memory (LSTM) to predict the price of six common indices, including Shanghai Composite Index, Shenzhen Component Index, Dow Jones Index, Nasdaq Index, Nikkei 225 and S&P 500. The model contains two paths of CNN and one path of LSTM to extract features. In our experiment with 10 years historic data of six indexes, the proposed CNN-LS achieved MSE of 0.5994 and MAE of 0.5427 on the testing set, both of which are better than MAE and MSE of five recent methods for stock prediction.","PeriodicalId":131256,"journal":{"name":"2021 14th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 14th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CISP-BMEI53629.2021.9624337","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
Stock market is one of the most important parts in the financial market. Numerous time series forecasting methods have been developed for predicting the stock price. Feature extraction is essential to many of these forecasting models. Highly related features can improve the accuracy of the forecasting model. This paper proposes a new model named CNN-LS that combines Convolution Neural Networks (CNN) with Long Short-Term Memory (LSTM) to predict the price of six common indices, including Shanghai Composite Index, Shenzhen Component Index, Dow Jones Index, Nasdaq Index, Nikkei 225 and S&P 500. The model contains two paths of CNN and one path of LSTM to extract features. In our experiment with 10 years historic data of six indexes, the proposed CNN-LS achieved MSE of 0.5994 and MAE of 0.5427 on the testing set, both of which are better than MAE and MSE of five recent methods for stock prediction.