Copper price forecasted by hybrid neural network with Bayesian Optimization and wavelet transform

Liu, Kailei; Cheng, Jinhua; Yi, Jiahui

doi:10.1016/j.resourpol.2021.102520

Cited by 35 publications

(16 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is crucial for analyzing nonlinear and non-stationary economic and financial time series, which can interact differently on different time scales [26][27][28][29][30][31][32][33][34][35]. In connection with such undoubted advantages, methods for forecasting nonlinear non-stationary economic and financial time series based on wavelet packet transform and combined methods have recently been actively developed, including Wavelet Artificial Neural Networks (WANN), Wavelet Least-Squares Support Vector Machine (WLSSVM), and Multivariate Adaptive Regression Splines (MARS) [36][37][38][39][40][41][42][43][44][45][46]. Their results indicate a significant increase in the performance and accuracy of traditional time series forecasting models in combination with wavelet packet transform (WPT).…”

Section: Literature Reviewmentioning

confidence: 99%

Improving the Accuracy of Forecasting the TSA Daily Budgetary Fund Balance Based on Wavelet Packet Transforms

Karaev

Gorlova

Sedova

et al. 2022

Journal of Open Innovation: Technology, Market, and Complexity

View full text Add to dashboard Cite

Improving the accuracy of cash flow forecasting in the TSA is the key to fulfilling government payment obligations, minimizing the cost of maintaining the cash reserve, providing the absence of outstanding debt accumulation, and ensuring investment in various financial instruments to obtain additional income. The article describes a method for improving the accuracy of forecasting a time series composed of daily budgetary fund balances in the TSA, based on its preliminary decomposition using a discrete wavelet packet transform of the Daubechies family. This makes it possible to increase the accuracy of traditional forecasting methods from 80% to more than 96%. The decomposition level varied from one to eight to minimize the mean absolute error and improve the forecasting accuracy. Calculations of statistical tests for adequacy confirm the effectiveness of the proposed method for improving forecasting accuracy. The scientific novelty of the proposed method for improving the forecasting accuracy of time series from daily budgetary fund balances in the TSA lies in proving the need for preliminary timeseries decomposition and subsequent construction of forecasts for the obtained parts, resulting in high forecasting accuracy. The result differs significantly from traditional econometric methods (ARIMA/SARIMA), characterized by a much lower accuracy (50–80%) and a decrease in forecasting accuracy with an increase in the forecast horizon. This article is novel, as it forms a new approach to solving the problem of increasing the efficiency of using budgetary funds, associated with improving the accuracy of forecasting daily budgetary fund balance in the TSA.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Improving the Accuracy of Forecasting the TSA Daily Budgetary Fund Balance Based on Wavelet Packet Transforms

Karaev

Gorlova

Sedova

et al. 2022

Journal of Open Innovation: Technology, Market, and Complexity

View full text Add to dashboard Cite

show abstract

“…The Bayesian framework was originally applied to the optimization of parameters of a neural network model, with good results [22]. The Bayesian evidence framework was used to obtain the optimal parameter values by maximizing parameter distribution [23].…”

Section: Algorithm Optimizationmentioning

confidence: 99%

Identification of Near Geographical Origin of Wolfberries by a Combination of Hyperspectral Imaging and Multi-Task Residual Fully Convolutional Network

Cui

Hao

et al. 2022

Foods

View full text Add to dashboard Cite

Ningxia wolfberry is the only wolfberry product with medicinal value in China. However, the nutritional elements, active ingredients, and economic value of the wolfberry vary considerably among different origins in Ningxia. It is difficult to determine the origin of wolfberry by traditional methods due to the same variety, similar origins, and external characteristics. In the study, we have for the first time used a multi-task residual fully convolutional network (MRes-FCN) under Bayesian optimized architecture for imaging from visible-near-infrared (Vis-NIR, 400–1000 nm) and near-infrared (NIR-1700 nm) hyperspectral imaging (HSI) technology to establish a classification model for near geographic origin of Ningxia wolfberries (Zhongning, Guyuan, Tongxin, and Huinong). The denoising auto-encoder (DAE) was used to generate augmented data, then principal component analysis (PCA) was combined with gray level co-occurrence matrix (GLCM) to extract the texture features. Finally, three datasets (HSI, DAE, and texture) were added to the multi-task model. The reshaped data were up-sampled using transposed convolution. After data-sparse processing, the backbone network was imported to train the model. The results showed that the MRes-FCN model exhibited excellent performance, with the accuracies of the full spectrum and optimum characteristic spectrum of 95.54% and 96.43%, respectively. This study has demonstrated that the MRes-FCN model based on Bayesian optimization and DAE data augmentation strategy may be used to identify the near geographical origin of wolfberries.

show abstract

“…Grid search is inefficient even with parallel computing. Past work has shown that random search is more efficient than grid search [32], but the problem is that it is easy to miss accurate solutions. In contrast to traditional methods, Bayesian optimization generates every guess based on previous training results.…”

Section: Bayesian Optimizationmentioning

confidence: 99%

“…This process is quite difficult computationally due to the long execution time of a tailored architecture. Thus, this paper proposes a method to automatically produce accurate predictions of PV power without manually tuning the deep learning architectures [31,32]. Zhou et al [33] developed a PV forecasting framework with a signal decomposition technique and a multi-objective chameleon swarm algorithm to predict short-term PV power.…”

Section: Introductionmentioning

confidence: 99%

Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods

Guo

Yan

2022

Sensors

View full text Add to dashboard Cite

The accurate prediction of photovoltaic (PV) power is essential for planning power systems and constructing intelligent grids. However, this has become difficult due to the intermittency and instability of PV power data. This paper introduces a deep learning framework based on 7.5 min-ahead and 15 min-ahead approaches to predict short-term PV power. Specifically, we propose a hybrid model based on singular spectrum analysis (SSA) and bidirectional long short-term memory (BiLSTM) networks with the Bayesian optimization (BO) algorithm. To begin, the SSA decomposes the PV power series into several sub-signals. Then, the BO algorithm automatically adjusts hyperparameters for the deep neural network architecture. Following that, parallel BiLSTM networks predict the value of each component. Finally, the prediction of the sub-signals is summed to generate the final prediction results. The performance of the proposed model is investigated using two datasets collected from real-world rooftop stations in eastern China. The 7.5 min-ahead predictions generated by the proposed model can reduce up to 380.51% error, and the 15 min-ahead predictions decrease by up to 296.01% error. The experimental results demonstrate the superiority of the proposed model in comparison to other forecasting methods.

show abstract

Copper price forecasted by hybrid neural network with Bayesian Optimization and wavelet transform

Cited by 35 publications

References 29 publications

Improving the Accuracy of Forecasting the TSA Daily Budgetary Fund Balance Based on Wavelet Packet Transforms

Improving the Accuracy of Forecasting the TSA Daily Budgetary Fund Balance Based on Wavelet Packet Transforms

Identification of Near Geographical Origin of Wolfberries by a Combination of Hyperspectral Imaging and Multi-Task Residual Fully Convolutional Network

Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods

Contact Info

Product

Resources

About