Search Results (177)

This paper presents an intelligent predictive system designed to support real-time decision making in the control of rubber blend mixing processes. The core of the system is a General Regression Neural Network (GRNN), which accurately predicts key process parameters, such as viscosity (expressed as torque), temperature, and energy consumption across varying masses of the processed material. The model can evaluate the mixing progress based on the initial 10% of input data, allowing early intervention and process optimisation. Experimental validation was conducted using a Brabender Plastograph EC Plus with a natural rubber-based blend in the mass range of 60–75 g. The GRNN kernel width parameter (σ) was optimised through a 10-fold cross-validation. High predictive accuracy was confirmed by values of the coefficient of determination (R²) approaching 1, and consistently low values of the root mean square error (RMSE). This system offers a robust and scalable solution for intelligent process control, productivity enhancement, and quality assurance across diverse industrial applications, beyond rubber blending. Full article

Metal-organic frameworks (MOFs) have been extensively studied for hydrogen storage due to their unique properties. This paper aims to develop several regression-based machine learning models to predict the hydrogen storage capacity of MOFs, including artificial neuron network (ANN), support vector regression (SVR), random forest (RF), extreme learning machine (ELM), kernel extreme learning machine (KELM), and generalized regression neural network (GRNN). An improved population-based metaheuristic optimization algorithm, the artificial lemming algorithm (ALA), is employed to select the hyperparameters of these machine learning models, enhancing their performance. All developed models are trained and tested using experimental data from multiple studies. The performance of the models is evaluated using various statistical metrics, complemented by regression plots, error analysis, and Taylor graphs to further identify the most effective predictive model. The results show that the ALA-RF model obtains the best performance in predicting hydrogen storage, with optimal values of coefficient of determination (R²), root mean square error (RMSE), Willmott’s index (WI), and weighted average percentage error (WAPE) in both training and testing phases (0.9845 and 0.9840, 0.2719 and 0.2828, 0.9961 and 0.9959, and 0.0667 and 0.0714, respectively). Additionally, pressure is identified as the most significant feature for predicting hydrogen storage in MOFs. These findings provide an intelligent solution for the selection of MOFs and optimization of operational conditions in hydrogen storage processes. Full article

The internal response mechanism of vegetation change in fragile high-altitude ecosystems is pivotal for ecological stability. This study focuses on the Lhasa River Basin (LRB) on the Tibetan Plateau (TP), a typical high-altitude fragile ecosystem where vegetation dynamics are highly sensitive to climate change and human activities. Utilizing MODIS surface reflectance data (MOD09Q1), a general regression neural network (GRNN) was applied to create a 250 m resolution fractional vegetation cover (FVC) dataset from 2001 to 2022, whose accuracy was verified with field survey data. Through methods like the Theil–Sen Median trend analysis, Mann–Kendall significance test, Hurst exponent, and geographical detector, the collaborative mechanism of 14 driving factors was systematically explored. Key conclusions are as follows: (1) The FVC in the LRB evolved in stages, first decreasing and then increasing, with 46.71% of the basin area expected to show an improvement trend in the future. (2) Among natural factors, elevation (q = 0.480), annual mean potential evapotranspiration (q = 0.362), and annual mean temperature (q = 0.361) are the main determinants of FVC spatiotemporal variation. (3) In terms of human activities, land use type has the highest explanatory power (q = 0.365) for FVC. (4) The interaction of two factors on FVC is stronger than that of a single factor, with the elevation–land use interaction being the most significant (q = 0.558). These results deepen our understanding of the interactions among vegetation, climate, and humans in fragile high-altitude ecosystems and provide a scientific basis for formulating zoned restoration strategies on the TP. Full article

Wholesale market prices of agricultural products, being essential to the daily lives of consumers, are closely tied to living standards and the overall stability of the agricultural market. The use of a single model to predict nonlinear and dynamic agricultural price time series often results in low accuracy due to suboptimal use of available information. To address this issue, this paper proposes a combined residual correction-based prediction method. Initially, the sparrow search algorithm (SSA) is used to optimize the penalty factors and kernel parameters of support vector regression (SVR) and the input weights and hidden layer biases of the extreme learning machine (ELM), thereby improving the convergence rate and predictive accuracy of these models. Subsequently, the induced ordered weighted averaging (IOWA) operator is applied to determine the weight vectors for the SSA-SVR and SSA-ELM models, reducing the fluctuating prediction accuracies of individual models at different times. Finally, the residuals of the generalized regression neural network (GRNN) model are forecasted using a combined residual correction method that integrates SSA-SVR and SSA-ELM based on the IOWA operator, refining the GRNN’s forecast outcomes. An empirical analysis was performed by comparing the results of nine individual forecasting models on monthly pork prices in Beijing. The findings indicate that the SSA-SVR, SSA-GRNN, and SSA-ELM models outperformed the SVR, GRNN, and ELM models in terms of forecasting accuracy, respectively. This improvement is attributed to the parameter optimization of the SVR, GRNN, and ELM models through the SSA. The proposed model also showed superior forecasting accuracy compared to the nine individual models. The results confirm that the proposed model is an effective tool for predicting agricultural product prices and can be applied to forecast prices of other agricultural products with similar characteristics. Full article

In this article, we introduce a novel deep learning hybrid model that integrates attention Transformer and gated recurrent unit (GRU) architectures to improve the accuracy of cryptocurrency price predictions. By combining the Transformer’s strength in capturing long-range patterns with GRU’s ability to model short-term and sequential trends, the hybrid model provides a well-rounded approach to time series forecasting. We apply the model to predict the daily closing prices of Bitcoin and Ethereum based on historical data that include past prices, trading volumes, and the Fear and Greed Index. We evaluate the performance of our proposed model by comparing it with four other machine learning models, two are non-sequential feedforward models: radial basis function network (RBFN) and general regression neural network (GRNN), and two are bidirectional sequential memory-based models: bidirectional long short-term memory (BiLSTM) and bidirectional gated recurrent unit (BiGRU). The model’s performance is assessed using several metrics, including mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE), along with statistical validation through the non-parametric Friedman test followed by a post hoc Wilcoxon signed-rank test. Results demonstrate that our hybrid model consistently achieves superior accuracy, highlighting its effectiveness for financial prediction tasks. These findings provide valuable insights for enhancing real-time decision making in cryptocurrency markets and support the growing use of hybrid deep learning models in financial analytics. Full article

Rapid urbanization in China is intensifying travel demand while making transport the nation’s third-largest source of carbon emissions. Anticipating continued growth in private-car fleets, this study integrates vehicle-stock forecasting with multi-scenario emission modeling to identify effective decarbonization pathways for Chinese cities. First, Kendall rank and grey relational analyses are combined to screen the key drivers of car ownership, creating a concise input set for prediction. A Lévy-flight-enhanced Sparrow Search Algorithm (LSSA) is then used to optimize the smoothing factor of the Generalized Regression Neural Network (GRNN), producing the Levy flight-improved Sparrow Search Algorithm optimized Generalized Regression Neural Network (LSSA-GRNN) model for annual fleet projections. Second, a three-tier scenario framework—Baseline, Moderate Low-Carbon, and Enhanced Low-Carbon—is constructed in the Long-range Energy Alternatives Planning System (LEAP) platform. Using Ningbo as a case study, the LSSA-GRNN outperforms both the benchmark Sparrow Search Algorithm optimized Generalized Regression Neural Network (SSA-GRNN) and the conventional GRNN across all accuracy metrics. Results indicate that Ningbo’s car fleet will keep expanding to 2030, albeit at a slowing rate. Relative to 2022 levels, the Enhanced Low-Carbon scenario delivers the largest emission reduction, driven primarily by accelerated electrification, whereas public transport optimization exhibits a slower cumulative effect. The methodological framework offers a transferable tool for cities seeking to link fleet dynamics with emission scenarios and to design robust low-carbon transport policies. Full article

The location of housing has a significant influence on its pricing. Generally, spatial self-correlation and spatial heterogeneity phenomena affect housing price data. Additionally, time is a crucial factor in housing price modeling, as it helps understand market trends and fluctuations. Currency market fluctuations also directly affect housing prices. Therefore, in addition to the physical features of the property, such as the area of the residential unit and building age, the rate of exchange (dollar price) is added to the independent variable set. This study used the real estate transaction records from Iran’s registration system, covering February, May, August, and November in 2017–2019. Initially, 7464 transactions were collected, but after preprocessing, the dataset was refined to 7161 records. Unlike feedforward neural networks, the generalized regression neural network does not converge to local minimums, so in this research, the Geographically, Temporally, and Characteristically Weighted Generalized Regression Neural Network (GTCW-GRNN) for housing price modeling was developed. In addition to being able to model the spatial–time heterogeneity available in observations, this algorithm is accurate and faster than MLR, GWR, GRNN, and GCW-GRNN. The average index of the adjusted coefficient of determination in other methods, including the MLR, GWR, GTWR, GRNN, GCW-GRNN, and the proposed GTCW-GRNN in different modes of using Euclidean or travel distance and fixed or adaptive kernel was equal to 0.760, 0.797, 0.854, 0.777, 0.774, and 0.813, respectively, which showed the success of the proposed GTCW-GRNN algorithm. The results showed the importance of the variable of the dollar and the area of housing significantly. Full article

Interior permanent magnet synchronous motors (IPMSMs) are critical for electric vehicle traction and industrial systems, yet optimizing their performance under high-dimensional design spaces remains computationally challenging. This study proposes a surrogate-assisted multi-objective optimization framework tailored for limited sample sizes. The methodology integrates random forest (RF) and analysis of variance (ANOVA) for variable importance analysis to reduce model complexity, followed by a Generalized Regression Neural Network (GRNN) to establish an efficient surrogate model. A multi-objective particle swarm optimization (MOPSO) algorithm generates Pareto-optimal solutions, while an entropy-weighted distance metric objectively selects the final design. Experimental results demonstrate that the optimized IPMSM achieves a 4.62% increase in average output torque, a 0.15% improvement in efficiency, and a 10.48% reduction in torque ripple compared to the prototype. Finite element analysis validates the consistency between predicted and simulated outcomes, with relative errors below 2.92%. The framework effectively balances computational efficiency and accuracy, offering a data-driven approach for motor optimization under constrained experimental resources. Full article

Solar radiation is one of the most abundant energy sources in the world and is a crucial parameter that must be researched and developed for the sustainable projects of future generations. This study evaluates the performance of different machine learning methods for solar radiation prediction in Konya, Turkey, a region with high solar energy potential. The analysis is based on hydro-meteorological data collected from NASA/POWER, covering the period from 1 January 1984 to 31 December 2022. The study compares the performance of Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Gated Recurrent Unit (GRU), Bidirectional GRU (Bi-GRU), LSBoost, XGBoost, Bagging, Random Forest (RF), General Regression Neural Network (GRNN), Support Vector Machines (SVM), and Artificial Neural Networks (MLANN, RBANN). The hydro-meteorological variables used include temperature, relative humidity, precipitation, and wind speed, while the target variable is solar radiation. The dataset was divided into 75% for training and 25% for testing. Performance evaluations were conducted using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the coefficient of determination (R²). The results indicate that LSTM and Bi-LSTM models performed best in the test phase, demonstrating the superiority of deep learning-based approaches for solar radiation prediction. Full article

The identification of the most pertinent site parameters to classify soils in terms of their amplification of seismic ground motions is still of prime interest to earthquake engineering and codes. This study investigates many options for improving soil classifications in order to reduce the deviation between “exact” predictions using wave propagation and the method used in seismic codes based on amplification (site) factors. To this end, an exhaustive parametric study is carried out to obtain nonlinear responses of sets of 324 clay and sand columns and to constitute the database for neuronal network methods used to predict the regression equations of the amplification factors in terms of seismic and site parameters. A wide variety of parameters and their combinations are considered in the study, namely, soil depth, shear wave velocity, the stiffness of the underlaying bedrock, and the intensity and frequency content of the seismic excitation. A database of AFs for 324 nonlinear soil profiles of sand and clay under multiple records with different intensities and frequency contents is obtained by wave propagation, where soil nonlinearity is accounted for through the equivalent linear model and an iterative procedure. Then, a Generalized Regression Neural Network (GRNN) is used on the obtained database to determine the most significant parameters affecting the AFs. A second neural network, the Radial Basis Function (RBF) network, is used to develop simple and practical prediction equations. Both the whole period range and specific short-, mid-, and long-period ranges associated with the AFs, F_a, F_v, and F_l, respectively, are considered. The results indicate that the amplification factor of an arbitrary soil profile can be satisfactorily approximated with a limited number of sites and the seismic record parameters (two to six). The best parameter pair is (PGA; resonance frequency, f₀), which leads to a standard deviation reduction of at least 65%. For improved performance, we propose the practical triplet $\left({PGA;V}_{s30};f_0\right)$ with V_s30 being the average shear wave velocity within the upper 30 m of soil below the foundation. Most other relevant results include the fact that the AFs for long periods (F_l) can be significantly higher than those for short or mid periods for soft soils. Finally, it is recommended to further refine this study by including additional soil parameters such as spatial configuration and by adopting more refined soil models. Full article

Visibility can be interpreted as the largest distance of an object that can be recognized or detected under a bright environment that can be used as an environmental indicator for weather conditions and air pollution. The accuracy of the classical approach of visibility calculation, in which meteorological laws and image feature extraction from digital images are used, depends on the quality and noise disturbances of the image. Therefore, artificial intelligence (AI) and digital image approaches have been proposed for visibility estimation in the past. Image features for the whole digital image are generated by pre-trained convolutional neural networks, and the Artificial Neural Network (ANN) is designed for correlation between image features and visibilities. Instead of using the information of the whole digital images, past research has been proposed to identify effective subregions from which image features are generated. A generalized regression neural network (GRNN) was designed to correlate the image features with the visibilities. Past research results showed that this method is more accurate than the classical approach of using handcrafted features. However, the selection of effective subregions of digital images is not fully automated and is based on manual selection by expert judgments. In this paper, we proposed an automatic effective subregion selection method using landmark object extraction techniques. Image features are generated from these LMO subregions, and the ANN is designed to approximate the mapping between LMO regions’ feature values and visibility values. The experimental results show that this approach can minimize the reductant information for ANN training and improve the accuracy of visibility estimation as compared to the single image approach. Full article

In this paper, we investigate the particle deposition distribution characteristics in granular fertilizer spreading, establish a relationship model between operational parameters and particle deposition distribution, and design an unmanned aerial vehicle (UAV) fertilizer particle deposition prediction system based on neural network decision making, which provides a decision-making basis for the variable fertilizer application model under multifactorial interactions. The particle deposition distribution data under different operating parameters were obtained by EDEM simulation and data superposition methods, and a generalized regression neural network (GRNN) based on a genetic algorithm (GA) was used to establish the prediction model of particle deposition, which was validated by bench test. The results show that the prediction accuracy and training effect of the GA-GRNN model are better than those of the GRNN, with a coefficient of determination of 0.839, and that the results of the GA-GRNN model are closer to the actual data when predicting the effective amplitude of the deposition amount, which is more accurate. The bench-scale validation test shows that the simulation is basically consistent with the actual measured deposition amount, and the deposition curve is normally distributed with a lateral error of about 3%. The results validate the reliability of the data superposition method for particle deposition distribution and the feasibility of the GA-GRNN model in multifactor prediction, which provides a theoretical basis and practical guidance for precision fertilizer application operations using agricultural UAVs. Full article

The fluctuations in solar irradiance and temperature throughout the year require an accurate methodology for forecasting the generated current of a PV system based on its specifications. The optimal technique must effectively manage rapid weather fluctuations while maintaining high accuracy in forecasting the performance of a PV panel. This work presents a comparative examination of supervised learning algorithms optimized with particle swarm optimization for estimating photovoltaic output current. The empirical formula’s measured currents are compared with outputs from various neural networks techniques, including feedforward neural networks (FFNNs), the general regression network known as GRNN, cascade forward neural networks also known as CFNNs, and adaptive fuzzy inference systems known as ANFISs, all optimized for enhanced accuracy using the particle swarm optimization (PSO) method. The ground data utilized for these models comprises hourly irradiations and temperatures from 2023, sourced from several places in South Africa. The accuracy levels indicated by statistical error margins from the root mean square error (RMSE), mean bias error (MBE), and mean absolute percentage error (MAPE) imply a universal enhancement in the algorithms’ precision upon optimization. Full article

The current study aims to model the South African crude oil prices using the hybrid of Box-Jenkins autoregressive integrated moving average (ARIMA) and Neural Networks (NNs). This study introduces a hybrid approach to forecasting methods aimed at resolving the issues of lack of precision in forecasting. The proposed methodology includes two models, namely, hybridisation of ARIMA with artificial neural network (ANN)-based Extreme Learning Machine (ELM) and ARIMA with general regression neural network (GRNN) to model both linear and nonlinear simultaneously. The models were compared with the base ARIMA model. The study utilised monthly time series data spanning from January 2021 to March 2023. The formal stationarity test confirmed that the crude oil price series is integrated of order one, I(1). For the linear process, the ARIMA (2,1,2) model was identified as the best fit for the series and successfully passed all diagnostic tests. The ARIMA-ANN-based ELM hybrid model outperformed both the individual ARIMA model and the ARIMA-GRNN hybrid. However, the ARIMA model also showed better performance than the ARIMA-GRNN hybrid, highlighting its strong competitiveness compared to the ARIMA-ANN-based ELM model. The hybrid models are recommended for use by policy makers and practitioners in general. Full article

In order to solve the problems of frequent pressure fluctuations caused by frequent action of the unloading valve of the pump station and serious hydraulic shock due to the variable amount of fluid used in the hydraulic support system of the coal mining face and the irregularity of the load suffered by the system, a pump–valve cooperative multi-mode stabilizing control method based on a digital unloading valve was proposed. Firstly, a prototype of a digital unloading valve under high-pressure and high water-based conditions was developed, and a digital control scheme was proposed to control the pilot valve by a servo motor to adjust the system pressure in real time. Then, an experimental platform for simulating the hydraulic bracket and a co-simulation model was constructed, and the validity of the co-simulation model was verified through experiments. Secondly, a collaborative multi-mode pressure stabilization control method for the pump valve based on a GRNN (General Regression Neural Network) was established to control the flow and pressure output of the emulsion pumping station according to the actual working conditions. Finally, numerical research and experimental verification were carried out for different working conditions to prove the effectiveness of this method. The results showed that the proposed pressure stabilization control method could adaptively adjust the working state of the digital unloading valve and the liquid supply flow of the emulsion pump station according to the working condition of the hydraulic support, effectively reducing the frequency and amplitude of the system pressure fluctuations and making the system pressure more stable. Full article