High-Performance Time Series Prediction With Predictive Error Compensated Wavelet Neural Networks

Üstündağ, Burak Berk; Kulaglic, Ajla

doi:10.1109/access.2020.3038724

Cited by 29 publications

(17 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 5 shows the modeling errors of different recent methods. In order to do a fair comparison, we use the same models as the Arima model and PEC-WNN in [ 47 ]. We form the problems of prediction, modeling and identification into the same sense: we use the trained model to estimate the next value in the time series, and the objects are the same.…”

Section: Simulationsmentioning

confidence: 99%

Complex Valued Deep Neural Networks for Nonlinear System Modeling

López-Pacheco

2021

Neural Process Lett

View full text Add to dashboard Cite

Deep learning models, such as convolutional neural networks (CNN), have been successfully applied in pattern recognition and system identification recent years. But for the cases of missing data and big noises, CNN does not work well for dynamic system modeling. In this paper, complex valued convolution neural network (CVCNN) is presented for modeling nonlinear systems with large uncertainties. Novel training methods are proposed for CVCNN. Comparisons with other classical neural networks are made to show the advantages of the proposed methods.

show abstract

Section: Simulationsmentioning

confidence: 99%

Complex Valued Deep Neural Networks for Nonlinear System Modeling

López-Pacheco

2021

Neural Process Lett

View full text Add to dashboard Cite

show abstract

“…The Mackey-Glass system is presented as a model of white blood cell production [12]. Let us first predict the time series data generated from a time-delayed Mackey-Glass nonlinear differential equation, which has been used as a benchmark by the researchers for validating their prediction algorithms [13], [14]. The series can be defined as mentioned in [2].…”

Section: A Mackey-glass Time Series As Predicted By Enf-adbel Networkmentioning

confidence: 99%

Expanded Neo-Fuzzy Adaptive Decayed Brain Emotional Learning Network for Online Time Series Predication

Milad

2021

IEEE Access

View full text Add to dashboard Cite

The Neo-Fuzzy integrated Adaptive Decayed Brain Emotional Learning (NF-ADBEL) network has recently been proposed for online time series predicting problems. The NF-ADBEL network is suitable for online time series prediction with shorter update intervals and offers features such as fast learning, accuracy, simplicity, and lower computational complexity. However, the neo-fuzzy neuron network in NF-ADBEL was integrated only in the orbitofrontal cortex (OFC) part of the ADBEL network. This paper aims to further improve the performance of the NF-ADBEL network by integrating the neo-fuzzy neuron network into the amygdala (AMY) section as well, inspired by a fully integrated version of a neo-fuzzy-based pattern recognizer. As is known, the AMY has two outputs: one response is based on imprecise information received from the thalamus, and the second response is based on information received from the sensory cortex. In this study, the imprecise response generation is operated as previously, while the other AMY process is treated by neo-fuzzy neurons. The resultant network is called Expanded Neo-Fuzzy integrated Adaptive Decayed Brain Emotional Learning (ENF-ADBEL). The modified network is still simple and meets the requirement for online prediction problems. A few chaotic and stochastic nonlinear systems, namely the Mackey-Glass, Lorenz, Rossler, disturbance storm time index, Narendra dynamic plant identification, wind speed and wind power series, are used to evaluate the performance of the proposed network in terms of the root mean squared error (RMSE) and correlation coefficient (COR) criteria in a MATLAB programming environment.INDEX TERMS Amygdala, Brain emotional decayed learning, Chaotic time series, Disturbance storm time index, Dynamic plant identification, Forecasting, MATLAB, Neo-fuzzy network, Wind speed, wind power series.

show abstract

“…Instead of applying the error data patterns back into the same network through the increased amount of inputs and nodes, we propose using an additional NN that is trained by the error of the first NN. Specifically, when the WT of the input data patterns and the error data patterns are used in this method (PEC-WNN) overall accuracy significantly raises while the time complexity remains less than the unified network equivalent of the solution [37]. This proposed data efficiency raising strategy can then be extended by using more amount of NNs trained by the error pattern of the superposed prediction and additional data can also be fused.…”

Section: Predictive Error Compensation Wavelet Neural Network Modelmentioning

confidence: 99%

Stock Price Prediction Using Predictive Error Compensation Wavelet Neural Networks

Kulaglic¹,

Üstündağ²

2021

Computers, Materials &Amp; Continua

Self Cite

View full text Add to dashboard Cite

Machine Learning (ML) algorithms have been widely used for financial time series prediction and trading through bots. In this work, we propose a Predictive Error Compensated Wavelet Neural Network (PEC-WNN) ML model that improves the prediction of next day closing prices. In the proposed model we use multiple neural networks where the first one uses the closing stock prices from multiple-scale time-domain inputs. An additional network is used for error estimation to compensate and reduce the prediction error of the main network instead of using recurrence. The performance of the proposed model is evaluated using six different stock data samples in the New York stock exchange. The results have demonstrated significant improvement in forecasting accuracy in all cases when the second network is used in accordance with the first one by adding the outputs. The RMSE error is 33% improved when the proposed PEC-WNN model is used compared to the Long Short-Term Memory (LSTM) model. Furthermore, through the analysis of training mechanisms, we found that using the updated training the performance of the proposed model is improved. The contribution of this study is the applicability of simultaneously different time frames as inputs. Cascading the predictive error compensation not only reduces the error rate but also helps in avoiding overfitting problems.

show abstract

High-Performance Time Series Prediction With Predictive Error Compensated Wavelet Neural Networks

Cited by 29 publications

References 30 publications

Complex Valued Deep Neural Networks for Nonlinear System Modeling

Complex Valued Deep Neural Networks for Nonlinear System Modeling

Expanded Neo-Fuzzy Adaptive Decayed Brain Emotional Learning Network for Online Time Series Predication

Stock Price Prediction Using Predictive Error Compensation Wavelet Neural Networks

Contact Info

Product

Resources

About