Abstract
Due to its complexity, financial time-series forecasting is regarded as one of the most challenging problems. During the past two decades, nonlinear modeling techniques, such as artificial neural networks, were commonly employed to solve a variety of time-series problems. Recently, however, deep neural network has been found to be more efficient than those in many application domains. In this article, we propose a model based on deep neural networks that improves the forecasting of stock prices. We investigate the impact of combining deep learning techniques with multiresolution analysis to improve the forecasting accuracy. Our proposed model is based on an empirical wavelet transform which is shown to outperform traditional stationary wavelet transform in capturing price fluctuations at different time scales. The proposed model is demonstrated to be substantially more effective than other models when evaluated on the S&P500 stock index and Mackey-Glass time series.
Highlights
A time-series is a sequence of data values taken at spaced successive points in time
The proposed approach transformed the financial time-series and extracted wavelet coefficients by Auto-correlation Shell Representation (ASR) and applied Bayesian method of automatic relevance determination (ARD) to select best features for the first layer that is composed of multiple Multilayer Perceptron (MLP) predictors
The results show that the combination of Convolutional Neural Network (CNN) and Multiresolution Analysis (MRA) outperforms the compared deep learning networks (DNNs)
Summary
A time-series is a sequence of data values taken at spaced successive points in time. The proposed approach transformed the financial time-series and extracted wavelet coefficients by ASR and applied Bayesian method of automatic relevance determination (ARD) to select best features for the first layer that is composed of multiple MLP predictors. Few researchers proposed combining wavelet transforms with deep learning techniques to learn financial time-series forecasting model. Persio and Honchar [47] investigated the merits of applying wavelet analysis with deep learning techniques in financial time-series forecasting They conducted many experiments comparing the forecasting performance of LSTM and CNN. The applied methodology is based on using the reconstructed data from the wavelet analysis They compared the proposed wavelet LSTM network with many other models, trained on the same data, and they found it to be superior. The proposed methodology was evaluated by conducting experiments for both short- and long-term forecasting using two benchmark datasets, namely S&P500 dataset and Mackey-glass time series
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.
