Modeling reference evapotranspiration using extreme learning machine and generalized regression neural network only with temperature data

Stream-flow forecasting using extreme learning machines: A case study in a semi-arid region in Iraq

Air overpressure (AOp) is a hazardous effect induced by the blasting method in surface mines. Therefore, it needs to be predicted to reduce the potential risk of damage. The aim of this study is to offer an efficient method to predict AOp using a cascaded forward neural network (CFNN) trained by Levenberg–Marquardt (LM) algorithm, called the CFNN-LM model. Additionally, a generalized regression neural network (GRNN) and extreme learning machine (ELM) were employed to demonstrate the accuracy level of the proposed CFNN-LM model. To conduct the CFNN-LM, GRNN, and ELM models, an extensive database, related to four quarry sites in Malaysia, was used including 62 sets of dependent and independent parameters. Next, the performances of the aforementioned models were checked and discussed through statistical criteria and efficient graphical tools. Finally, the results showed the superiority of CFNN-LM (R2 = 0.9263 and RMSE = 3.0444) over GRNN (R2 = 0.7787 and RMSE = 5.1211) and ELM (R2 = 0.6984 and RMSE = 6.2537) models in terms of prediction accuracy. Furthermore, three different regression analysis metrics were used to perform the sensitivity analysis, and according to the obtained results, the maximum charge per delay ( $$\beta$$ = 0.475, SE = 0.115, t-test = 4.125) was considered as the most influential feature in modeling the AOp.

This paper presents the development of a general regression neural network (GRNN) model for the prediction of annual municipal solid waste (MSW) generation at the national level for 44 countries of different size, population and economic development level. Proper modelling of MSW generation is essential for the planning of MSW management system as well as for the simulation of various environmental impact scenarios. The main objective of this work was to examine the potential influence of economy crisis (global or local) on the forecast of MSW generation obtained by the GRNN model. The existence of the so-called structural breaks that occur because of the economic crisis in the studied period (2000-2012) for each country was determined and confirmed using the Chow test and Quandt-Andrews test. Two GRNN models, one which did not take into account the influence of the economic crisis (GRNN) and another one which did (SB-GRNN), were developed. The novelty of the applied method is that it uses broadly available social, economic and demographic indicators and indicators of sustainability, together with GRNN and structural break testing for the prediction of MSW generation at the national level. The obtained results demonstrate that the SB-GRNN model provide more accurate predictions than the model which neglected structural breaks, with a mean absolute percentage error (MAPE) of 4.0% compared to 6.7% generated by the GRNN model. The proposed model enhanced with structural breaks can be a viable alternative for a more accurate prediction of MSW generation at the national level, especially for developing countries for which a lack of MSW data is notable.

Water consumption of plants is a key parameter for formulating irrigation system, and the precise prediction play a important role in improving the use efficiency of limited water resources. In this experiment, by using the method of artificial neural network and MATLAB DATA PROCESSING SYSTEM combined with the meteorological data of air temperature, relative air humidity, solar radiation, wind speed, soil water content and dew point temperature as the input variable, the author established the artificial neural network system to forecast the seedling water consumption of P.×euramericana cv."74/76", and through the experiments it has been examined that two neural network system models can be applied in forecasting water consumption of seedlings, and the average relative error of Back Propagation (BP) neural network prediction model was 0.07, the General Regression Neural Network (GRNN) prediction model was 0.05, moreover, the latter had good stability, while that of the former was poor. Therefore, we propose that GRNN model can be used in prediction of seedling water consumption. Furthermore, the maximum relative error of GRNN predication model was 0.106, the minimum relative error was 0.015. The GRNN model is superior to the BP neural network model that the former performs a higher forecasting accuracy with relatively shorter time consumption and faster speed in training.

Modeling of energy consumption and related GHG (greenhouse gas) intensity and emissions in Europe using general regression neural networks

Forecasting chlorine residuals in a water distribution system using a general regression neural network

Real-time porosity prediction using gas-while-drilling data and machine learning with reservoir associated gas: Case study for Hassi Messaoud field, Algeria

Rapid detection of food quality indicators using ELM and near-infrared spectroscopy.

Wind speed forecasting is critical for a variety of engineering tasks, such as wind energy harvesting, scheduling of a wind power system, and dynamic control of structures (e.g., wind turbine, bridge, and building). Wind speed, which has characteristics of random, nonlinear and uncertainty, is difficult to forecast. Nowadays, machine learning approaches (generalized regression neural network (GRNN), back propagation neural network (BPNN), and extreme learning machine (ELM)) are widely used for wind speed forecasting. In this study, two schemes are proposed to improve the forecasting performance of machine learning approaches. One is that optimization algorithms, i.e., cross validation (CV), genetic algorithm (GA), and particle swarm optimization (PSO), are used to automatically find the optimal model parameters. The other is that the combination of different machine learning methods is proposed by finite mixture (FM) method. Specifically, CV-GRNN, GA-BPNN, PSO-ELM belong to optimization algorithm-assisted machine learning approaches, and FM is a hybrid machine learning approach consisting of GRNN, BPNN, and ELM. The effectiveness of these machine learning methods in wind speed forecasting are fully investigated by one-year field monitoring data, and their performance is comprehensively compared.

At present, there are mainly three research ways on the distribution of snow load: field measurement, experimental simulation, and numerical simulation. They all need expensive equipment, specimen model used only once, and slight defective simulation software. Artificial Neural Networks (ANNs) rely on a large number of input functions to estimate unknown or known occurrences. This paper applies ANNs to predict the uneven distribution of snow loads. First, the snow load database is collected and set up, including all the data from field measurement and experimental simulation. Second, part of the data from the 1 m3 model is used as a training set to train Backpropagation Neural Networks (BPNN) and Generalized Regression Neural Networks (GRNN). The parameters of the two neural networks are determined (such as the smoothing factor in the GRNN), and the rest of the data is used to examine the model; the prediction and measurement results of the two models are compared. Finally, based on different wind direction angles and wind speed, a GRNN model is established to predict the uneven snow load distribution. Through comparative analysis, GRNN is more suitable for predicting the uneven distribution of snow loads, especially in predicting snow thickness. In particular, the prediction effect of the model on the snow load distribution trend is good, and the prediction ability of the neural network model is displayed.

To enhance technical support for monitoring the condition and managing the overall health of marine diesel engines, this study introduces a baseline model for marine diesel engine exhaust gas temperature (EGT). This model integrates the Improved Slime Mould Algorithm (ISMA) with the Generalized Regression Neural Network (GRNN) for real-time EGT monitoring. For the experimental dataset, this study introduces methods such as Mahalanobis distance, sliding window filtering algorithm, and Pearson's algorithm to establish a comprehensive data processing workflow. Subsequently, the ISMA algorithm is employed to optimize the hyperparameter (σ) value in the GRNN, and the optimized σ value is then input into the GRNN model to establish the optimal baseline model for EGT prediction. Validation is conducted using historical ’8L20’ diesel engine data. Experimental results demonstrate that the ISMA-GRNN baseline model leverages the respective advantages of both methods and exhibits superior performance in prediction accuracy, convergence speed, and robustness compared to RNN, BiLSTM, GRNN, GA-GRNN, and PSO-GRNN models. The model effectively utilizes historical monitoring data to deeply extract internal information, enabling accurate baseline model establishment of diesel engine exhaust gas temperature. This not only contributes to the realization of condition monitoring and early warning functions for ship engines, improving ship energy efficiency and environmental performance, but also provides a new approach for intelligent operation and maintenance of ships.

Enhancing oil recovery using in-situ combustion (ISC) is an attractive alternative, especially for heavy crudes. During ISC, part of the hydrocarbon is pyrolysed/oxidised, which generates heat and deposits fuel in the combustion front. In this study, crude reactions during ISC are modelled after their thermogravimetry thermo-oxidative profiles using advanced machine learning systems. The model inputs include the weight per cent of asphaltenes, resins, and °API gravity of the oil as well as the heating rate and the temperature. Four types of artificial neural networks (ANNs); namely multilayer perceptron (MLP), generalised regression neural network (GRNN), cascade-forward neural network (CFNN), and radial basis function (RBF) neural network, were employed to develop models for accurate prediction of the weight per cent of residual crude oil based on 2289 experimental data points. Moreover, three optimisation algorithms; including Bayesian Regularisation (BR), Levenberg–Marquardt (LM), and Scaled Conjugate Gradient (SCG) were applied in the training step of MLP and CFNN to improve the prediction ability. GRNN provided the most accurate prediction with ∼2.3% overall average absolute per cent relative error and coefficient of determination of 0.9983. GRNN model is reliable for crude oils with °API gravity of 5–35 and up to 820°C. Lastly, a mathematical correlation was developed to estimate the residual crude oil from thermogravimetry analysis using gene expression programming (GEP). GEP also predicted the thermo-oxidative profile with high accuracy. On the basis of sensitivity analysis, residue formation during crude oil oxidation was impacted the most by the temperature, oil °API gravity, and asphaltenes content, respectively. The Leverage approach identified 2.9% of the data points as doubtful.

Calibration of solar radiation models for Europe using Meteosat Second Generation and weather station data

House price forecasting with neural networks

Solar photovoltaics (PV) has advanced at an unprecedented rate and the global cumulative installed PV capacity is growing exponentially. However, the ability to forecast PV power remains a key technical challenge due to the variability and uncertainty of solar irradiance resulting from the changes of clouds. Ground-based remote sensing with high temporal and spatial resolution may have potential for solar irradiation forecasting, especially under cloudy conditions. To this end, we established two ultra-short-term forecasting models of global horizonal irradiance (GHI) using Ternary Linear Regression (TLR) and Back Propagation Neural Network (BPN), respectively, based on the observation of a ground-based sky imager (TSI-880, Total Sky Imager) and a radiometer at a PV plant in Dunhuang, China. Sky images taken every 1 min (minute) were processed to determine the distribution of clouds with different optical depths (thick, thin) for generating a two-dimensional cloud map. To obtain the forecasted cloud map, the Particle Image Velocity (PIV) method was applied to the two consecutive images and the cloud map was advected to the future. Further, different types of cloud fraction combined with clear sky index derived from the GHI of clear sky conditions were used as the inputs of the two forecasting models. Limited validation on 4 partly cloudy days showed that the average relative root mean square error (rRMSE) of the 4 days ranged from 5% to 36% based on the TLR model and ranged from 12% to 32% based on the BPN model. The forecasting performance of the BPN model was better than the TLR model and the forecasting errors increased with the increase in lead time.