Search

The random fluctuation of solar energy poses significant challenges for direct solar utilization systems. Solar-driven membrane reactors (SMMR) are efficient devices for energy conversion and pure hydrogen production, featuring thermal inertia and complex physicochemical processes. To realize advance control of SMMR under solar variations to promote their practical application, accurate performance prediction is essential. This study develops a deep learning-assisted performance prediction model for SMMR using Bayesian optimized long short-term memory (BO-LSTM) network and weather classification. Time-series datasets with similar environmental features are generated through SMMR multi-physics modeling and trained by BO-LSTM. The BO-LSTM demonstrates high accuracy in both short-time (1-day) and long-time (40-day) predictions, with a correlation coefficient exceeding 0.997 and a deviation of ±0.0016 in long-term predictions. Weather classification ensures consistent prediction accuracy across different weather. Compared to RF, SVM, BPNN, and CNN models, BO-LSTM provides smoother prediction curves and significantly improves accuracy and stability, reducing RMSE by 44.6% on sunny days and 53.3% on cloudy days in short-time predictions, and by 56.7% and 68.0% in long-time predictions, respectively. The BO-LSTM proposed in this study accurately predicts SMMR performance under various weather, guiding practical applications and offering a reference for prediction and control in solar thermal utilization systems. [Display omitted] • Self-tuning BO-LSTM is used to predict performances of solar-driven membrane reactor. • Weather classification achieves equally excellent prediction for sunny and cloudy. • Errors of long-term data training is smaller than the short-term and R2 is 0.997. • Mean prediction errors of BO-LSTM is reduced by 45%–75% compared to other models. • Predictive accuracy and stability of BO-LSTM for easy integration in control systems. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Zhengzhou University (Natural Science Edition) is the property of Journal of Zhengzhou University (Natural Science Edition) Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Time series analysis and prediction have attained significant attention from the research community in the past few decades. However, the prediction accuracy of the models highly depends on the models' learning process. In order to optimize resource usage, a better learning methodology, in terms of accuracy and learning time, is needed. In this context, the current research work proposes EvoLearn, a novel method to improve and optimize the learning process of neural-based models. The presented technique integrates the genetic algorithm with back-propagation to train model weights during the learning process. The fundamental idea behind the proposed work is to select the best components from multiple models during the training process to obtain an adequate model. To demonstrate the applicability of EvoLearn, the method is tested on the state-of-the-art neural models (namely MLP, DNN, CNN, RNN, and GRU), and performances are compared. Furthermore, the presented study aims to forecast two types of time series, i.e. air pollution and energy consumption time series, using the developed framework. In addition, the considered neural models are tested on two datasets of each time series type. From the performance comparison and evaluation of EvoLearn using a one-tailed paired T-test against the conventional back-propagation-based learning approach, it was found that the proposed method significantly improves the prediction accuracy. [ABSTRACT FROM AUTHOR]

Stage-built long-span bridges deform with temperature, affecting alignment to design needs. In this paper, a model for predicting temperature time series is proposed, which can predict temperatures in engineering practice and utilize the predicted results to adjust the elevation of stage construction. The model employs convolutional neural networks (CNNs) for initial feature extraction, followed by bidirectional long short-term memory (BiLSTM) layers to capture temporal dependencies. An attention mechanism is applied to the LSTM output, enhancing the model's ability to focus on the most relevant parts of the sequence. The Crested Porcupine Optimizer (CPO) is used to fine-tune parameters like the number of LSTM units, dropout rate, and learning rate. The experiments on the measured temperature data of an under-construction cable-stayed bridge are conducted to validate our model. The results indicate that our model outperforms the other five models in comparison, with all the R2 values exceeding 0.97. The average of the mean absolute error (MAE) on the 30 measure points is 0.19095, and the average of the root mean square error (RMSE) is 0.28283. Furthermore, the model's low sensitivity to data makes it adaptable and effective for predicting temperatures and adjusting the elevation in large-span bridge construction. [ABSTRACT FROM AUTHOR]

Mine flooding accidents have occurred frequently in recent years, and the predicting of mine water inflow is one of the most crucial flood warning indicators. Further, the mine water inflow is characterized by non-linearity and instability, making it difficult to predict. Accordingly, we propose a time series prediction model based on the fusion of the Transformer algorithm, which relies on self-attention, and the LSTM algorithm, which captures long-term dependencies. In this paper, Baotailong mine water inflow in Heilongjiang Province is used as sample data, and the sample data is divided into different ratios of the training set and test set in order to obtain optimal prediction results. In this study, we demonstrate that the LSTM-Transformer model exhibits the highest training accuracy when the ratio is 7:3. To improve the efficiency of search, the combination of random search and Bayesian optimization is used to determine the network model parameters and regularization parameters. Finally, in order to verify the accuracy of the LSTM-Transformer model, the LSTM-Transformer model is compared with LSTM, CNN, Transformer and CNN–LSTM models. The results prove that LSTM-Transformer has the highest prediction accuracy, and all the indicators of its model are well improved. [ABSTRACT FROM AUTHOR]

The capacity of lithium-ion batteries gradually degrades over time, presenting unforeseen risks and losses. Models based on data-driven approaches and neural network predictions can offer early warnings for battery failure. However, common models often face challenges with error accumulation in predicting future capacity changes, and insufficient data complicates model training. To address this, a novel method for predicting remaining battery life is proposed. This method incorporates IC analysis, DTV analysis, and Local Linear Embedding algorithms for efficient feature extraction and uses TimeGAN to augment training data, creating an integrated prediction framework. It combines the long-sequence prediction capabilities of the Pyraformer network with the dynamic multi-variable relationship capturing of the BiLSTM network, enhancing the model's understanding of capacity degradation trends. Comparative experiments demonstrate that this approach achieves higher prediction accuracy than traditional simple neural networks. Additionally, ablation studies further confirm the effectiveness of the introduced techniques in prediction tasks. [ABSTRACT FROM AUTHOR]

The establishment of the NOx concentration prediction model is a prerequisite for the selective catalytic reduction (SCR) flue gas denitriflcation systems in thermal power plants to overcome measurement delays in continuous emission monitoring systems and achieve precise control. This paper proposes a NOx emission prediction method for thermal power plants based on the attention mechanism and the gated recurrent unit (Attention-GRU) to address the issues of low accuracy and cumbersome feature selection process in NOx emission prediction models established for the SCR system under complex working conditions. Firstly, this article utilizes the GRU to extract data features of NOx emissions from thermal power plants at the time scale, in order to establish a long-term prediction model. Secondly, the feature dimension clustering the k-means algorithm can effectively improve the learning efficiency of temporal features and reduce the computational complexity of subsequent algorithms. Finally, the parameter attention mechanism is introduced to autonomously select favorable temporal features for predicting NOx emissions in thermal power plants, replacing the complex and time-consuming data screening process. This paper has also collected historical datasets of the SCR system form a thermal power plant under complex and simple working conditions to verify the effectiveness of Attention-GRU. The experimental results show that compared with existing methods (see, e.g., ELM, RF, SVM, BPNN, LSTM, GRU, BiLSTM, and CNN-LSTM), our method has higher prediction accuracy of NOx emissions from thermal power plants, which helps to improve the control performance of SCR systems to reduce atmospheric pollution. At the same time, this article has also found that the attention mechanism can significantly reduce the dependence of existing methods on the feature selection process, effectively solve the interference problem of noise data on model feature extraction, and further improve the prediction accuracy of NOx emissions in thermal power plants. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Graphics is the property of Journal of Graphics Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this paper we investigate the use of reinforcement-learning based prediction approaches for a real drinking-water treatment plant. Developing such a prediction system is a critical step on the path to optimizing and automating water treatment. Before that, there are many questions to answer about the predictability of the data, suitable neural network architectures, how to overcome partial observability and more. We first describe this dataset, and highlight challenges with seasonality, nonstationarity, partial observability, and heterogeneity across sensors and operation modes of the plant. We then describe General Value Function (GVF) predictions—discounted cumulative sums of observations–and highlight why they might be preferable to classical n-step predictions common in time series prediction. We discuss how to use offline data to appropriately pre-train our temporal difference learning (TD) agents that learn these GVF predictions, including how to select hyperparameters for online fine-tuning in deployment. We find that the TD-prediction agent obtains an overall lower normalized mean-squared error than the n-step prediction agent. Finally, we show the importance of learning in deployment, by comparing a TD agent trained purely offline with no online updating to a TD agent that learns online. This final result is one of the first to motivate the importance of adapting predictions in real-time, for non-stationary high-volume systems in the real world. [ABSTRACT FROM AUTHOR]

The expected healthcare (HC) inflation rate (IR) (HCIR) is an important variable for all economic agents within HC systems. In recent years, during the COVID‐19 pandemic, Iran has experienced a high HCIR in its health system. In this context, a robust approximation of HCIR will be a helpful tool for health authorities and other decision makers. Using monthly time series data of HCIR in Iran, we developed various forecasting techniques based on classical smoothing methods, decomposition ETS (error, trend, and seasonality) approaches, autoregressive (AR) integrated moving average (ARIMA), seasonal ARIMA (SARIMA), and a multilayer nonlinear AR artificial neural network (NARANN) with several training algorithms including Bayesian regularization (BR), Levenberg–Marquardt (LM), scaled conjugate gradient (SCG), Broyden–Fletcher–Goldfarb–Shanno (BFGS) quasi‐Newton, conjugate gradient with Powell–Beale restarts (CGB), conjugate gradient with Fletcher–Reeves updates (CGF), and resilient propagation (RPROP) algorithms. Initially, based upon various criteria and possible combinations, we selected the superior model for each method separately. After that, the best model in each category is involved in 6‐ and 12‐multi‐step‐ahead prediction. In this stage, several error criteria are calculated. According to our findings, in a six‐step forecasting window, the Holt–Winters with a multiplicative seasonal pattern and SARIMA showed less bias, though compared to other alternatives like NARANN‐lm/br, the difference was relatively small. In the next process, by doubling the forecasting window, it is observed that artificial neural network (ANN) (i.e., Bayesian NARANN) strictly outperformed other models. As a result, in shorter steps, the Holt–Winters method can provide a better prediction, while in longer windows, Bayesian NARANN can be implemented vigorously for the prediction. Finally, we used 10 models to predict the future trend of HCIR in Iran till the end of July 2024. [ABSTRACT FROM AUTHOR]

Accurate prediction of air quality index is a challenging task, in order to solve the gradient problem of traditional neural network methods in the time series prediction process as well as to improve the prediction accuracy, the study proposes a hybrid quantum neural network prediction model based on quantum activation function. The model utilizes a quantum classical convolutional neural network to tap into spatial correlations between different time periods and combines it with a quantum activation function so as to better avoid the gradient problem and solve the death RELU problem for better spatial feature extraction, and then uses the long short term memory neural network to capture the observations at different times. Experiments were conducted on different air quality datasets using the model, which proved that the proposed quantum activation function optimized hybrid quantum neural network algorithm showed more remarkable advantages in prediction accuracy than other model algorithms. [ABSTRACT FROM AUTHOR]

Echo state network (ESN) has widely attracted many researchers due to its training process without backpropagation. However, it is hard for single ESN to fit those complex and polytrophic situations. Under this case, a novel multiscale deep ESN (MDESN) is developed in this study, which integrates the deep learning, parallel structure and multiscale reservoir state matrix mode, respectively. The deep framework is responsible for collecting multivariate dimensional reservoir states; Upgradation of various reservoir states could efficiently dig hidden characteristics for decoding. The parallel structure could be trained simultaneously to decrease time consumption through the multiple thread method. According to the multivariate high dimensional mapping method, MDESN could obtain more robust generalization information compared to the state-of-the-art models. MDESN is evaluated in two chaos prediction benchmarks (Lorenz and Mackey-Glass) and real solar irradiance predictions. Various prediction horizons including one-step-ahead and multi-step-ahead prediction are designed and conducted respectively to verify the effectiveness and robustness of MDESN. Furthermore, RMSE, MAE, MAPE, and R are adopted to evaluate our proposed model. The statistical results demonstrate that MDESN has the best-performing adaptability and robustness compared to classical ESN, deep ESN (DESN), and other state-of-the-art models. [ABSTRACT FROM AUTHOR]

The prediction of power equipment activity plays a vital role in optimizing power resource dispatch, ensuring supply and demand balance, and guiding network planning and management. However, due to the complex nonlinear, multi-scale, and multivariate characteristics of power grid communication networks service activity (PCNSA) data, it is often challenging to capture its intrinsic patterns and dynamic changes through a simple model. To address this issue, this paper proposes a power grid communication network service activity prediction method based on convolutional neural network, bidirectional gated recurrent unit, and attention mechanism, referred to as CBGA. CNN is used to extract features from time-series data, BiGRU is used to capture long-term feature changes in the data, and the attention mechanism is used to enhance the extraction of important information. To validate the performance of the proposed method, experiments and comparisons were conducted on three real power grid communication network datasets, demonstrating that our method shows better performance and robustness. [ABSTRACT FROM AUTHOR]

Accurately estimating the Remaining Useful Life (RUL) of lithium-ion batteries is crucial for maintaining the safe and stable operation of rechargeable battery management systems. However, this task is often challenging due to the complex temporal dynamics. Recently, attention-based networks, such as Transformers and Informer, have been the popular architecture in time series forecasting. Despite their effectiveness, these models with abundant parameters necessitate substantial training time to unravel temporal patterns. To tackle these challenges, we propose a straightforward MLP-Mixer-based architecture named "Intra–Inter Patch Mixer" (IIP-Mixer), which leverages the strengths of multilayer perceptron (MLP) models to capture both local and global temporal patterns in time series data. Specifically, it extracts information using an MLP and performs mixing operations along both intra-patch and inter-patch dimensions for battery RUL prediction. The proposed IIP-Mixer comprises parallel dual-head mixer layers: the intra-patch mixing MLP, capturing local temporal patterns in the short-term period, and the inter-patch mixing MLP, capturing global temporal patterns in the long-term period. Notably, to address the varying importance of features in RUL prediction, we introduce a weighted loss function in the MLP-Mixer-based architecture, marking the first time such an approach has been employed. Our experiments demonstrate that IIP-Mixer achieves competitive performance in battery RUL prediction, outperforming other popular time series frameworks, such as Informer and DLinear, with relative reductions in mean absolute error (MAE) of 24% and 10%, respectively. [ABSTRACT FROM AUTHOR]

Bluetooth Low Energy Mesh (BLE Mesh) enables Bluetooth flexibility and coverage by introducing Low-Power Nodes (LPNs) and enhanced networking protocol. It is also a commonly used communication method in sensor networks. In BLE Mesh, LPNs are periodically woken to exchange messages in a stop-and-wait way, where the tradeoff between energy and efficiency is a hard problem. Related works have reduced the energy consumption of LPNs mainly in the direction of changing the bearer layer, improving time synchronization and broadcast channel utilization. These algorithms improve communication efficiency; however, they cause energy loss, especially for the LPNs. In this paper, we propose a constrained flooding algorithm based on time series prediction and lightweight GBN (Go-Back-N). On the one hand, the wake-up cycle of the LPNs is determined by the time series prediction of the surrounding load. On the other, LPNs exchange messages through lightweight GBN, which improves the window and ACK mechanisms. Simulation results validate the effectiveness of the Time series Prediction and LlightWeight GBN (TP-LW) algorithm in energy consumption and throughput. Compared with the original algorithm of BLE Mesh, when fewer packets are transmitted, the throughput is increased by 214.71%, and the energy consumption is reduced by 65.14%. [ABSTRACT FROM AUTHOR]

The warning of the potential faults occurring in the future in a sucker rod well can help technicians adjust production strategies in time. It is of great significance for safety during well production. In this paper, the key characteristic parameters of dynamometer cards were predicted by a temporal neural network to implement the warning of different working conditions which might result in failures. First, a one-dimensional damped-wave equation was used to eliminate the dynamic loads' effect of surface dynamometer cards by converting them into down-hole dynamometer cards. Based on the down-hole dynamometer cards, the characteristic parameters were extracted, including the load change, the position of the valve opening and closing point, the dynamometer card area, and so on. The mapping relationship between the characteristic parameters and working conditions (classification model) was obtained by the Xgboost algorithm. Meanwhile, the noise in these parameters was reduced by wavelet transformation, and the rationality of the results was verified. Second, the Encoder–Decoder and multi-head attention structures were used to set up the time series prediction model. Then, the characteristic parameters were predicted in a sequence-to-sequence way by using historical characteristic parameters, date, and pumping parameters as input. At last, by inputting the predicted results into the classification model, a working conditions warning method was created. The results showed that noise reduction improved the prediction accuracy significantly. The prediction relative error of most characteristic parameters was less than 15% after noise reduction. In most working conditions, their F1 values were more than 85%. Most Recall values could be restored to over 90% of those calculated by real parameters, indicating few false negative cases. In general, the warning method proposed in this paper can predict faulty working conditions that may occur in the future in a timely manner. [ABSTRACT FROM AUTHOR]

Dissolved oxygen is an important index to measure water quality. The accurate prediction of dissolved oxygen can provide reference for scientific management and protection of water environment. Considering that dissolved oxygen was affected by many external factors and the data had strong nonlinear and non-stationary characteristics, the DA-CNN-BiLSTM dissolved oxygen prediction model was proposed. The CNN layer was used to extract local features of data. Spatial attention focused on features that had a higher impact on prediction outcomes. BiLSTM mined the attribute relationships of the input sequence. Temporal attention layer captured the time dependence of different moments. The model was applied to the dissolved oxygen prediction of three water quality monitoring stations in Minjiang River, Fujian Province. By comparing with the baseline model, it shows that the DA-CNN-BiLSTM model has a better prediction of DO concentration compared with the baseline model. The predicted value of the model is closer to the measured value and the DO prediction performance is optimal. The prediction performance of the model has been improved after adding the spatial attention mechanism. [ABSTRACT FROM AUTHOR]

[Objective] The dynamic prediction of carbon emission from cold chain distribution is an important basis for the accurate assessment of carbon emission and its green credit grade. Facing the fact that the carbon emission of vehicles is affected by multiple factors, such as road condition information, driving characteristics, refrigeration parameters, etc., a dynamic prediction model of carbon emission was proposed from refrigerated vehicles that integrates multi-source information. [Methods] The backbone feature extraction network, neck feature fusion network and loss function of YOLOv8s was firstly improved. The full-dimensional dynamic convolution was introduced into the backbone feature extraction network, and the multidimensional at‐ tention mechanism was introduced to capture the contextual key information to improve the model feature extraction capability. A pro‐ gressive feature pyramid network was introduced into the feature extraction network, which reduced the loss of key information by fusing features layer by layer and improved the feature fusion efficiency. The road condition information recognition model based on improved YOLOv8s was constructed to characterize the road condition information in terms of the number of road vehicles and the percentage of pixel area. Pearson's correlation coefficient was used to compare and analyze the correlation between carbon emissions of refrigerated vehicles and different influencing factors, and to verify the necessity and criticality of the selection of input parameters of the carbon emission prediction model. Then the iTransformer temporal prediction model was improved, and the external attention mechanism was introduced to enhance the feature extraction ability of iTransformer model and reduce the computational complexity. The dynamic prediction model of carbon emission of refrigerated vehicles based on the improved iTransformer was constructed by taking the road condition information, driving characteristics (speed, acceleration), cargo weight, and refrigeration parameters (temper‐ ature, power) as inputs. Finally, the model was compared and analyzed with other models to verify the robustness of the road condi‐ tion information and the prediction accuracy of the vehicle carbon emission dynamic prediction model, respectively. [Results and Discussions] The results of correlation analysis showed that the vehicle driving parameters were the main factor affecting the intensity of vehicle carbon emissions, with a correlation of 0.841. The second factor was cargo weight, with a correlation of 0.807, which had a strong positive correlation. Compared with the vehicle refrigeration parameters, the road condition information had a stronger correlation between vehicle carbon emissions, the correlation between refrigeration parameters and the vehicle carbon emis‐ sions impact factor were above 0.67. In order to further ensure the accuracy of the vehicle carbon emissions prediction model, The pa‐ per was selected as the input parameters for the carbon emissions prediction model. The improved YOLOv8s road information recog‐ nition model achieved 98.1%, 95.5%, and 98.4% in precision, recall, and average recognition accuracy, which were 1.2%, 3.7%, and 0.2% higher than YOLOv8s, respectively, with the number of parameters and the amount of computation being reduced by 12.5% and 31.4%, and the speed of detection being increased by 5.4%. This was due to the cross-dimensional feature learning through full-dimen‐ sional dynamic convolution, which fully captured the key information and improved the feature extraction capability of the model, and through the progressive feature pyramid network after fusing the information between different classes through gradual step-bystep fusion, which fully retained the important feature information and improved the recognition accuracy of the model. The predictive performance of the improved iTransformer carbon emission prediction model was better than other time series prediction models, and its prediction curve was closest to the real carbon emission curve with the best fitting effect. The introduction of the external attention mechanism significantly improved the prediction accuracy, and its MSE, MAE, RMSE and R2 were 0.026 1 %VOL, 0.079 1 %VOL, 0.161 5 %VOL and 0.940 0, respectively, which were 0.4%, 15.3%, 8.7% and 1.3% lower, respectively, when compared with iTrans‐ former. As the degree of road congestion increased, the prediction accuracy of the constructed carbon emission prediction model increased. [Conclusions] The carbon emission prediction model for cold chain distribution under multi-source information fusion proposed in this study can realize accurate prediction of carbon emission from refrigerated vehicles, provide theoretical basis for rationally formulating carbon emission reduction strategies and promoting the development of low-carbon cold chain distribution. [ABSTRACT FROM AUTHOR]

The purpose of this study was to develop a machine learning model to reconstruct time series kinematic and kinetic profiles of the ankle and knee joint across six different tasks using an ankle-mounted IMU. Four male collegiate basketball players performed repeated tasks, including walking, jogging, running, sidestep cutting, max-height jumping, and stop-jumping, resulting in a total of 102 movements. Ankle and knee flexion-extension angles and moments were estimated using motion capture and inverse dynamics and considered 'actual data' for the purpose of model fitting. Synchronous acceleration and angular velocity data were collected from right ankle-mounted IMUs. A time-series feature extraction model was used to determine a set of features used as input to a random forest regression model to predict the ankle and knee kinematics and kinetics. Five-fold cross-validation was performed to verify the model accuracy, and statistical parametric mapping was used to determine the difference between the predicted and experimental time series. The random forest regression model predicted the time-series profiles of the ankle and knee flexion-extension angles and moments with high accuracy (Kinematics: R2 ranged from 0.782 to 0.962, RMSE ranged from 2.19° to 11.58°; Kinetics: R2 ranged from 0.711 to 0.966, RMSE ranged from 0.10 Nm/kg to 0.41 Nm/kg). There were differences between predicted and actual time series for the knee flexion-extension moment during stop-jumping and walking. An appropriately trained feature-based regression model can predict time series knee and ankle joint angles and moments across a wide range of tasks using a single ankle-mounted IMU. [ABSTRACT FROM AUTHOR]

Monitoring the condition of rotating machinery is critical in aerospace applications like aircraft engines and helicopter rotors. Faults in these components can lead to catastrophic outcomes, making early detection essential. This paper proposes a novel approach using vibration signals and time series prediction methods for fault detection in rotating aerospace machinery. By extracting relevant features from vibration signals and using prediction models, fault severity can be effectively quantified. Our experimental results show that the proposed method has potential in early fault detection and is applicable to various types of bearing faults and the different statuses of these faults under complex running conditions, achieving very good generalization ability. [ABSTRACT FROM AUTHOR]

To address the problem of undesirable prediction accuracy of a deep echo state network caused by redundant structures in the reservoirs, a pruning algorithm for the deep echo state network based on detrended multiple cross-correlation was proposed. Firstly, according to the detrended covariance function and the detrended variance function, the detrended cross-correlation coefficient between each two neurons in the selected reservoirs in turn was calculated, and the detrended cross-correlation matrix was constructed. Based on this matrix, the detrended multiple cross-correlation between a selected neuron and all remaining neurons in this reservoir could be evaluated. Subsequently, the connections from the highly correlated neurons in each reservoir to the output layer were pruned sequentially, thus removing redundant components in the network. Finally, the network after pruning was retrained by least squares regression to obtain the optimal deep echo state network topology. Simulation results showed that the prediction accuracy and memory capacity of the deep echo state network optimized by the proposed algorithm on Mackey-Glass time series were improved by 89. 80% and 30. 93%, respectively, and on Call time series by 14.34% and 0.10%, respectively. [ABSTRACT FROM AUTHOR]

The growing global requirement for food and the need for sustainable farming in an era of a changing climate and scarce resources have inspired substantial crop yield prediction research. Deep learning (DL) and machine learning (ML) models effectively deal with such challenges. This research paper comprehensively analyses recent advancements in crop yield prediction from January 2016 to March 2024. In addition, it analyses the effectiveness of various input parameters considered in crop yield prediction models. We conducted an in-depth search and gathered studies that employed crop modeling and AI-based methods to predict crop yield. The total number of articles reviewed for crop yield prediction using ML, meta-modeling (Crop models coupled with ML/DL), and DL-based prediction models and input parameter selection is 125. We conduct the research by setting up five objectives for this research and discussing them after analyzing the selected research papers. Each study is assessed based on the crop type, input parameters employed for prediction, the modeling techniques adopted, and the evaluation metrics used for estimating model performance. We also discuss the ethical and social impacts of AI on agriculture. However, various approaches presented in the scientific literature have delivered impressive predictions, they are complicated due to intricate, multifactorial influences on crop growth and the need for accurate data-driven models. Therefore, thorough research is required to deal with challenges in predicting agricultural output. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Frontiers of Computer Science & Technology is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract Time series analysis and prediction have attained significant attention from the research community in the past few decades. However, the prediction accuracy of the models highly depends on the models’ learning process. In order to optimize resource usage, a better learning methodology, in terms of accuracy and learning time, is needed. In this context, the current research work proposes EvoLearn, a novel method to improve and optimize the learning process of neural-based models. The presented technique integrates the genetic algorithm with back-propagation to train model weights during the learning process. The fundamental idea behind the proposed work is to select the best components from multiple models during the training process to obtain an adequate model. To demonstrate the applicability of EvoLearn, the method is tested on the state-of-the-art neural models (namely MLP, DNN, CNN, RNN, and GRU), and performances are compared. Furthermore, the presented study aims to forecast two types of time series, i.e. air pollution and energy consumption time series, using the developed framework. In addition, the considered neural models are tested on two datasets of each time series type. From the performance comparison and evaluation of EvoLearn using a one-tailed paired T-test against the conventional back-propagation-based learning approach, it was found that the proposed method significantly improves the prediction accuracy.

In this study, we proposed a novel modeling and prediction method employing both fractional calculus and the multi-deep assessment methodology (M-DAM), utilizing multifactor analysis across the entire dataset from 2000 to 2019 for comprehensive data modeling and prediction. We evaluated and reported the performance of M-DAM by modeling various economic factors such as current account balance (% of gross domestic product (GDP)), exports of goods and services (% of GDP), GDP growth (annual %), gross domestic savings (% of GDP), gross fixed capital formation (% of GDP), imports of goods and services (% of GDP), inflation (consumer prices, annual %), overnight interbank rate, and unemployment (total). The dataset used in this study covered the years between 2000 and 2019. The Group of Eight (G-8) countries and Turkey were chosen as the experimental domain. Furthermore, to understand the validity of M-DAM, we compared the modeling performance with multiple linear regression (MLR) and the one-step prediction performance with a recurrent neural network, long short-term memory (LSTM), and MLR. The results showed that in 75.04% of the predictions, M-DAM predicted the factors with less than 10% error. For the order of predictability considering the years 2018 and 2019, Germany was the most predictable country; the second group consisted of Canada, France, the UK, and the USA; the third group included Italy and Japan; and the fourth group comprised Russia. The least predictable country was found to be Turkey. Comparison with LSTM and MLR showed that the three methods behave complementarily.

Abstract Mine flooding accidents have occurred frequently in recent years, and the predicting of mine water inflow is one of the most crucial flood warning indicators. Further, the mine water inflow is characterized by non-linearity and instability, making it difficult to predict. Accordingly, we propose a time series prediction model based on the fusion of the Transformer algorithm, which relies on self-attention, and the LSTM algorithm, which captures long-term dependencies. In this paper, Baotailong mine water inflow in Heilongjiang Province is used as sample data, and the sample data is divided into different ratios of the training set and test set in order to obtain optimal prediction results. In this study, we demonstrate that the LSTM-Transformer model exhibits the highest training accuracy when the ratio is 7:3. To improve the efficiency of search, the combination of random search and Bayesian optimization is used to determine the network model parameters and regularization parameters. Finally, in order to verify the accuracy of the LSTM-Transformer model, the LSTM-Transformer model is compared with LSTM, CNN, Transformer and CNN–LSTM models. The results prove that LSTM-Transformer has the highest prediction accuracy, and all the indicators of its model are well improved.

ObjectiveThe dynamic prediction of carbon emission from cold chain distribution is an important basis for the accurate assessment of carbon emission and its green credit grade. Facing the fact that the carbon emission of vehicles is affected by multiple factors, such as road condition information, driving characteristics, refrigeration parameters, etc., a dynamic prediction model of carbon emission was proposed from refrigerated vehicles that integrates multi-source information.MethodsThe backbone feature extraction network, neck feature fusion network and loss function of YOLOv8s was firstly improved. The full-dimensional dynamic convolution was introduced into the backbone feature extraction network, and the multidimensional attention mechanism was introduced to capture the contextual key information to improve the model feature extraction capability. A progressive feature pyramid network was introduced into the feature extraction network, which reduced the loss of key information by fusing features layer by layer and improved the feature fusion efficiency. The road condition information recognition model based on improved YOLOv8s was constructed to characterize the road condition information in terms of the number of road vehicles and the percentage of pixel area. Pearson's correlation coefficient was used to compare and analyze the correlation between carbon emissions of refrigerated vehicles and different influencing factors, and to verify the necessity and criticality of the selection of input parameters of the carbon emission prediction model. Then the iTransformer temporal prediction model was improved, and the external attention mechanism was introduced to enhance the feature extraction ability of iTransformer model and reduce the computational complexity. The dynamic prediction model of carbon emission of refrigerated vehicles based on the improved iTransformer was constructed by taking the road condition information, driving characteristics (speed, acceleration), cargo weight, and refrigeration parameters (temperature, power) as inputs. Finally, the model was compared and analyzed with other models to verify the robustness of the road condition information and the prediction accuracy of the vehicle carbon emission dynamic prediction model, respectively.Results and DiscussionsThe results of correlation analysis showed that the vehicle driving parameters were the main factor affecting the intensity of vehicle carbon emissions, with a correlation of 0.841. The second factor was cargo weight, with a correlation of 0.807, which had a strong positive correlation. Compared with the vehicle refrigeration parameters, the road condition information had a stronger correlation between vehicle carbon emissions, the correlation between refrigeration parameters and the vehicle carbon emissions impact factor were above 0.67. In order to further ensure the accuracy of the vehicle carbon emissions prediction model, The paper was selected as the input parameters for the carbon emissions prediction model. The improved YOLOv8s road information recognition model achieved 98.1%, 95.5%, and 98.4% in precision, recall, and average recognition accuracy, which were 1.2%, 3.7%, and 0.2% higher than YOLOv8s, respectively, with the number of parameters and the amount of computation being reduced by 12.5% and 31.4%, and the speed of detection being increased by 5.4%. This was due to the cross-dimensional feature learning through full-dimensional dynamic convolution, which fully captured the key information and improved the feature extraction capability of the model, and through the progressive feature pyramid network after fusing the information between different classes through gradual step-by-step fusion, which fully retained the important feature information and improved the recognition accuracy of the model. The predictive performance of the improved iTransformer carbon emission prediction model was better than other time series prediction models, and its prediction curve was closest to the real carbon emission curve with the best fitting effect. The introduction of the external attention mechanism significantly improved the prediction accuracy, and its MSE, MAE, RMSE and R2 were 0.026 1 %VOL, 0.079 1 %VOL, 0.161 5 %VOL and 0.940 0, respectively, which were 0.4%, 15.3%, 8.7% and 1.3% lower, respectively, when compared with iTransformer. As the degree of road congestion increased, the prediction accuracy of the constructed carbon emission prediction model increased.ConclusionsThe carbon emission prediction model for cold chain distribution under multi-source information fusion proposed in this study can realize accurate prediction of carbon emission from refrigerated vehicles, provide theoretical basis for rationally formulating carbon emission reduction strategies and promoting the development of low-carbon cold chain distribution.

Stock price prediction has always been a focal topic in financial research and quantitative investment. Currently, most deep learning models for stock price prediction are based on batch learning settings, which require prior knowledge of the training dataset. These models are not scalable for real-time data stream prediction, and their performance decreases when the data distribution dynamically changes. To address the issue of poor prediction accuracy for non-stationary stock price data in existing research, this paper proposes an online stock price prediction model (Increformer) based on incremental learning and continuous attention mechanism. By leveraging continuous self-attention mechanism to capture the temporal dependencies among feature variables and employing continuous normalization mechanism to handle non-stationary data, the model enhances prediction accuracy through the incremental training strategy based on elastic weighting consolidation to acquire new knowledge from the data stream. Five public datasets are selected from the stock index and individual stock price sequences in the stock market for experiments. Experimental results demonstrate that Increformer effectively extracts temporal information and feature dimension correlation from the data, thus improving the prediction performance of stock prices. Additionally, ablation experiments are conducted to evaluate the effects and necessity of the continuous normalization mechanism, continuous attention mechanism, and incremental training strategy, thereby verifying the accuracy and generalizability of the proposed model. Increformer can effectively capture the trends and fluctuations of stock price series.

To fully utilize the potential of solar energy and effectively operate solar energy systems, it is vital to know solar radiation values. Modeling solar radiation data with high variability is a complex problem, and nonlinear methods are needed. This study proposes a hybrid model using CNN and LSTM architectures for solar radiation prediction. The performance and applicability of the proposed model are examined by comparing it with ARIMA statistical method and different machine learning methods such as Random Forest, Decision Tree and K-Nearest Neighbor. The study used hourly solar radiation values measured with a pyranometer on the Afyon Kocatepe University campus. As a result of the experimental studies, the CNNLSTM model is the most successful model in terms of RMSE, MAE, MAPE and r² evaluation criteria, achieving 16.92%, 17.25%, 17.24%, 22.89% more successful results than the Decision Trees model, which produced the most unsuccessful predictions, respectively. [ABSTRACT FROM AUTHOR]

Time series data, characterized by large volumes and wide-ranging applications, requires accurate predictions of future values based on historical data. Recent advancements in deep learning models, particularly in the field of time series forecasting, have shown promising results by leveraging neural networks to capture complex patterns and dependencies. However, existing models often overlook the influence of short-term cyclical patterns in the time series, leading to a lag in capturing changes and accurately tracking fluctuations in forecast data. To overcome this limitation, this paper introduces a new method that utilizes an interpolation technique to create a fine-scaled representation of the cyclical pattern, thereby alleviating the impact of the irregularity in the cyclical component and hence enhancing prediction accuracy. The proposed method is presented along with evaluation metrics and loss functions suitable for time series forecasting. Experiment results on benchmark datasets demonstrate the effectiveness of the proposed approach in effectively capturing cyclical patterns and improving prediction accuracy. [ABSTRACT FROM AUTHOR]

Abstract As sustainability emerges as a crucial factor in the development of modern enterprises, integrating environmental, social, and governance (ESG) information into financial assessments has become essential. ESG indicators serve as important metrics in evaluating a company’s sustainable practices and governance effectiveness, influencing investor trust and future growth potential, ultimately affecting stock prices. This study proposes an innovative approach that combines ESG sentiment index extracted from news with technical indicators to predict the S&P 500 index. By utilizing a deep learning model and exploring optimal window sizes, the study explores the best model through mean absolute percentage error (MAPE) as an evaluation metric. Additionally, an ablation test clarifies the influence of ESG and its causality with the S&P 500 index. The experimental results demonstrate improved predictive accuracy when considering ESG sentiment compared to relying solely on technical indicators or historical data. This comprehensive methodology enhances the advantage of stock price prediction by integrating technical indicators, which consider short-term fluctuations, with ESG information, providing long-term effects. Furthermore, it offers valuable insights for investors and financial market experts, validating the necessity to consider ESG for financial assets and introducing a new perspective to develop investment strategies and decision-making processes.

Aviation is one of the most global industries, and if we can model and predict a country’s air transportation flow and indicators ahead of time, we may be able to use it as a key decision-making tool for the management and operation process. This study proposes a new modeling, and prediction method that employs both fractional calculus and Multi Deep Assessment Methodology (MDAM) techniques. For the application, air passengers carried, air freight, available seat kilometers, number of flights, destination points, international travelers, international destination points, and international flight data between 2011 and 2019 for eight countries with the busiest airports were chosen. As a result, the highest modeling error was discovered to be Germany’s air transport freight factor expressed as a percentage of 1,59E-02. The percentage of predictions with errors less than 10% was 90.278. We also compared the performance of two different MDAM methodologies. The novel MDAM wd methodology proposed in this paper has a higher accuracy in aviation factors prediction and modeling.