Abstract
<p><em>During training process of LSTM, the prediction accuracy is affected by a variation of factors, including the selection of training samples, the network structure, the optimization algorithm, and the stock market status. This paper tries to conduct a systematic research on several influencing factors of LSTM training in context of time series prediction. The experiment uses Shanghai and Shenzhen 300 constituent stocks from 2006 to 2017 as samples. The influencing factors of the study include indicator sampling, sample length, network structure, optimization method, and data of the bull and bear market, and this experiment compared the effects of PCA, dropout, and L2 regularization on predict accuracy and efficiency. Indice sampling, number of samples, network structure, optimization techniques, and PCA are found to be have their scope of application. Further, dropout and L2 regularization are found positive to improve the accuracy. The experiments cover most of the factors, however have to be compared by data overseas. This paper is of significance for feature and parameter selection in LSTM training process.</em></p>
Highlights
In recent years, there have been many studies with LSTM, some of which especially suitable for stock market time series forecasting (Hochreiter & Schmidhuber, 1997)
During training process of LSTM, the prediction accuracy is affected by a variation of factors, including the selection of training samples, the network structure, the optimization algorithm, and the stock market status
The influencing factors of the study include indicator sampling, sample length, network structure, optimization method, and data of the bull and bear market, and this experiment compared the effects of PCA, dropout, and L2 regularization on predict accuracy and efficiency
Summary
There have been many studies with LSTM, some of which especially suitable for stock market time series forecasting (Hochreiter & Schmidhuber, 1997). The selection of various factors such as training samples, model structure, and optimization methods is often subjective. As a result, it has become the biggest problem in engineering applications. The research on the influencing factors of LSTM model prediction accuracy, includes sample characteristics, network structure selection and optimization methods. Maknickienė et al used LSTM in the USD/JPY exchange rate forecast (Maknickienė, Rutkauskas, & Maknickas, 2011) They found that neurons amounts and the number of training iterations were basically stable over a certain range. The above research is oriented to different fields, covering the training factors, network structure, neurons amounts, optimization methods and other influencing factors, but there are widespread problems such as insufficient training samples, insufficient factor selection, etc. Proposed techniques such as Dropout (Srivastava et al, 2014; Bluche, Kermorvant, & Louradour, 2015) and regularization (L2 regularization) (Theodoridis, 2015)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
