Gaussian Process Regression Algorithm Research Articles

Gasification process modelling and optimization using Gaussian process regression and hybrid population-based algorithms

Gasification process modelling and optimization using Gaussian process regression and hybrid population-based algorithms

Data-driven uncertainty quantification and sensitivity studies in free vibration behavior of bio-inspired helicoidal laminated composite cylindrical shells

The uncertainty in the ply orientation of bio-inspired helicoidal laminated composite cylindrical shells is quantified. The frequencies obtained from the higher-order shear deformation theory are trained to formulate a surrogate model in the multi-output Gaussian Process Regression algorithm framework, and the stochastic analysis is then carried out using the Monte Carlo simulation method. For simply supported and clamped shells, uncertainties in the top and bottom-most layers widely affect the free vibration behavior of the shells compared to the inner-most or central layers, whereas for cantilever shells, opposite trend is observed. The inclusion of noise widely affects the free vibration behavior.

Dual-indicators machine learning assisted processing high-quality laser-induced fluorine-doped graphene and its application on droplet velocity monitoring sensor

Laser-induced graphene (LIG) is widely used in electron devices owing to its characteristics of fine patterning and high precision. Doping is a commonly used approach to improve the quality of LIG. Thanks to the excellent superhydrophobicity and high electrical conductivity of laser-induced fluorine-doped graphene (F-LIG), it shows promising potential in developing electrodes for electron devices. However, it remains a major challenge to rapidly search for the optimal laser processing parameter space that can process high-quality F-LIG. In this work, a dual-indicators predictive machine learning (ML) method for optimizing the F-LIG process was proposed, utilizing a Gaussian process regression (GPR) algorithm. The mean square error (MSE) between the actual and predicted values of the data used for the model test was 0.00814, while the mean absolute percentage error (MAPE) was 10.33 %, demonstrating a certain degree of generalization without overfitting. Importantly, the F-LIG processed using parameters predicted by the ML model achieved excellent superhydrophobicity and conductivity simultaneously. Based on this high-quality F-LIG, a self-powered droplet velocity monitoring sensor was developed. This sensor can accurately work after 20,000 cycles, showing excellent stability. The proposed dual-indicator ML model assisted the optimization of the F-LIG process, which will provide a feasible and economical way to process high-quality graphene and related electron devices.

Predicting CBR values using gaussian process regression and meta-heuristic algorithms in geotechnical engineering

Predicting CBR values using gaussian process regression and meta-heuristic algorithms in geotechnical engineering

Erosion Prediction Gaussian Process Regression Algorithm for Alumina and Chromia Reinforced Nickel-Based High-Velocity Oxy-Fuel Coatings

Background:: Machine learning (ML) methodologies have demonstrated efficacy in the determination of erosion rates and material removal. In this context, a novel Erosion Prediction Gaussian Process Regression Algorithm (EPGPRA) was developed to predict the volumetric erosion in thermal spray coatings. Aim:: In this patent, a novel EPGPRA based model was developed to predict the volumetric loss of 30Al2O3 and 20Cr2O3 reinforced Ni-based coatings deposited using a high-velocity oxy-fuel (HVOF) process. Objective:: The objective of this patent is to develop a GPR model for the prediction of Ni-30Al2O3 and Ni-20Cr2O3 coatings. Methods:: Spraying powders were applied to the SS316L steel substrate in order to develop coatings. An erosion tester was used in order to investigate the wear resistance of HVOF-coated steel. The gathered experimental dataset is put to use in the construction of a powerful GPR model. The outcomes from GPR model were then measured against the values obtained from the experiments. To demonstrate the accuracy of the GPR model, the produced model is evaluated against various cutting-edge machine learning methods. Results:: This innovation was successful in terms of developing a new GPR model for wear prediction. The discrepancy between the actual and expected values is the smallest for Matern 5/2 (M5/2) GPR in the validation set. It was also lesser as compared to Ensemble Boosted Trees, Support Vector Machine, Linear regression, and Fine Tree. In terms of MSE, MAE, RMSE, and R2 the accuracy performance of the M5/2 GPR model was determined to be 9.8565×10-5, 0.0048884, 0.009928, and 0.93 correspondingly. Ni-Chromia coating performed better than the Ni-Alumina. Conclusion:: As per this patent, a novel EPGPRA-based model was developed, which is the better machine learning technique for wear prediction of Ni-based HVOF coatings.

Robust Gaussian process regression based on bias trimming

This paper presents a new robust Gaussian process regression (GPR) algorithm based on identifying and trimming outliers, and it can reduce the computation time and improve the accuracy compared with other robust algorithms. Currently, most existing robust GPR approaches model noise with heavy-tailed distributions that are less sensitive to outliers, such as the Laplace and Student-t distributions. However, a model that is designed this way will be analytically intractable, and the computationally expensive Bayesian approximation method needs to be adopted to obtain the posterior distribution. In our method, the deviation between normal points and outliers is modeled as a bias value, so the non-Gaussian likelihood approach, which consumes substantial time, is not necessary. Nevertheless, the decrease in calculation time is accompanied by a decrease in accuracy. To solve this problem, the trimming algorithm with an adaptive threshold is adopted to improve the prediction precision. Therefore, the proposed approach is simple but very computationally attractive. We conducted various simulation experiments to numerically evaluate the model with different outlier ratios and noise levels. We compared the approach with existing robust Gaussian process methods and other kinds of anomaly detection algorithms to demonstrate the excellent performance of the proposed method. Finally, industrial datasets are used to verify the method, and the results illustrate that it can significantly improve the practicality of the measurement system.

Investigations on the applicability of machine learning algorithms to optimize biodiesel composition for improved engine fuel properties

Selecting suitable biodiesel for the intended application is challenging due to the significant variations in the feedstock for producing biodiesel. The available models to predict biodiesel properties have limited applicability and reliability. The present work addresses these two challenges by developing reliable models based on machine learning algorithms for predicting engine fuel properties of biodiesel and optimizing biodiesel composition for better fuel properties. The models are developed using multilinear regression (MLR), artificial neural networks (ANN), support vector machine regression with grid search (SVMGS), Bayesian optimization (SVMBO) and grey-wolf optimization (SVMGWO) for hyperparameter tuning, Gaussian process regression (GPR), random forest (RF), and adaptive neuro-fuzzy inference system (ANFIS) algorithms. The models are trained to predict viscosity, cetane number, and calorific value from 13 methyl ester constituents of 70 biodiesels. SVMGS models predicted the viscosity, cetane number, and calorific value of 33 validation samples with a mean absolute percentage error of 1.54%, 1%, and 0.43%. Biodiesel composition was optimized to minimize viscosity and maximize cetane number and calorific value. The optimized composition exhibits 3.72 cSt viscosity, 57 cetane number, and 43 MJ/kg calorific value, which can be prepared by blending 68% ± 1% camelina and 32% ± 1% coconut oil. Applying machine learning algorithms to predict biodiesel properties yielded more accurate predictions than available models. It helped find the optimal composition for improved engine characteristics.

Optimization‐improved thermal–mechanical simulation of welding residual stresses in welded connections

AbstractThis paper presents a novel computer‐aided computational framework to determine the optimum shape parameters in a welding heat source model using a coupled supervised Gaussian process regression (GPR) and genetic algorithm (GA) approach in estimating the welding residual stresses. The experimental X‐ray diffraction (XRD) approach validates the optimization‐improved thermal–mechanical simulation method. The validation of the proposed approach includes the welded plates with simple geometry to welded pipe‐to‐plate and the circular hollow section joints with complex topologies. The optimization‐derived shape parameters in a welding heat source model lead to close estimates of the temperature history and the welding residual stresses for these types of welded joints considered in this study.

Fast flow field prediction of three-dimensional hypersonic vehicles using an improved Gaussian process regression algorithm

Conducting large-scale numerical computations to obtain flow field during the hypersonic vehicle engineering design phase can be excessively costly. Although deep learning algorithms enable rapid flow field prediction with high-precision, they require a significant investment in training samples, contradicting the motivation of reducing the cost of acquiring flow field. The combination of feature extraction algorithms and regression algorithms can also achieve high-precision prediction of flow fields, which is more suitable to tackle three-dimensional flow prediction with a small dataset. In this study, we propose a reduced-order model (ROM) for the three-dimensional hypersonic vehicle flow prediction utilizing proper orthogonal decomposition to extract representative features and Gaussian process regression with improved automatic kernel construction (AKC-GPR) to perform a nonlinear mapping of physical features for prediction. The selection of variables is based on sensitivity analysis and modal assurance criterion. The underlying relationship is unveiled between flow field variables and inflow conditions. The ROM exhibits high predictive accuracy, with mean absolute percentage error (MAPE) of total field less than 3.5%, when varying altitudes and Mach numbers. During angle of attack variations, the ROM only effectively reconstructs flow distribution by interpolation with a MAPE of 7.02%. The excellent small-sample fitting capability of our improved AKC-GPR algorithm is demonstrated by comparing with original AKC-GPRs with a maximum reduction in a MAPE of 35.28%. These promising findings suggest that the proposed ROM can serve as an effective approach for rapid and accurate vehicle flow predicting, enabling its application in engineering design analysis.

Potato Leaf Chlorophyll Content Estimation through Radiative Transfer Modeling and Active Learning

Leaf chlorophyll content (LCC) significantly correlates with crop growth conditions, nitrogen content, yield, etc. It is a crucial indicator for elucidating the senescence process of plants and can reflect their growth and nutrition status. This study was carried out based on a potato nitrogen and potassium fertilizer gradient experiment in the year 2022 at Keshan Farm, Qiqihar Branch of Heilongjiang Agricultural Reclamation Bureau. Leaf hyperspectral and leaf chlorophyll content data were collected at the potato tuber formation, tuber growth, and starch accumulation periods. The PROSPECT-4 radiative transfer model was employed to construct a look-up table (LUT) as a simulated data set. This was accomplished by simulating potato leaves’ spectral reflectance and chlorophyll content. Then, the active learning (AL) technique was used to select the most enlightening training samples from the LUT based on the measured potato data. The Gaussian process regression (GPR) algorithm was finally employed to construct the inversion models for the chlorophyll content of potato leaves for both the whole and single growth periods based on the training samples selected by the AL method and the ground measured data of the potatoes. The R2 values of model validation accuracy for the potato whole plantation period and three single growth periods are 0.742, 0.683, 0.828, and 0.533, respectively with RMSE values of 4.207, 4.364, 2.301, and 3.791 µg/cm2. Compared with the LCC inversion accuracy through LUT with a cost function, the validation accuracies of the GPR_PROSPECT-AL hybrid model were improved by 0.119, 0.200, 0.328, and 0.255, and the RMSE were reduced by 3.763, 2.759, 0.118, and 5.058 µg/cm2, respectively. The study results indicate that the hybrid method combined with the radiative transfer model and active learning can effectively select informative training samples from a data pool and improve the accuracy of potato LCC estimation, which provides a valid tool for accurately monitoring crop growth and growth health.

Machine learning assessment of hydrological model performance under localized water storage changes through downscaling

The coarse spatial resolution of the Gravity Recovery and Climate Experiment (GRACE) data has limited its application in the management of local-scale water resources. To address this limitation, we developed a new downscaling approach using predictors from regional and global hydrological models for a 15-year period (2002–2017) and tested it in the northern Great Artesian Basin, Australia. We used four different machine learning algorithms (support vector machine, partial least squares, gaussian process and random forest) to downscale the original GRACE estimate of 0.5° to a spatial grain size of 0.1° (global) and 0.05° (regional). This was based on precipitation, evapotranspiration and runoff estimates from the Famine Early Warning Systems Network Land Data Assimilation System (FLDAS) and Australian Water Outlook (AWO) hydrological models, respectively. The downscaled products were validated using 42 in-situ precipitation observations spread across the test region. We further evaluated which of the downscaled products best mimicked local-scale hydrology using a range of statistical metrics. Our results showed that regional hydrological models best characterized the dynamics of local scale hydrology (rainfall v. downscaled product), and the gaussian process regression algorithm made the best predictions for both models. The correlation coefficients for the raw values varied from 0.45 to 0.49 while that of the standardized values varied from 0.46 to 0.52 with the random forest model providing the best fitting for the regional-based products. The regional downscaling approach employed in this study may be readily integrated into local water resources planning programs.

Behavioral-level modelling of GaN HEMT large signal based on Pelican-Gaussian process regression algorithm

In this paper, a novel technology named Pelican-Gaussian process regression machine learning algorithm is proposed for modelling the large-signal characteristics of Gallium Nitride High Electron Mobility Transistors (GaN HEMT). Hyperparameter optimization in traditional Gaussian process regression algorithms tends to fall into local optimums and is overly dependent on the initial values. In order to solve this problem, the Pelican optimization algorithm is introduced to optimize the hyperparameters in Gaussian process regression algorithms in the article. The Pelican optimization algorithm is able to make the global exploration and local search ability of the algorithm be effectively balanced by helping particles to escape from the local optimal position. The I–V characteristics, output power, power gain, power gain efficiency and small-signal S-parameters of GaN HEMT devices are used to verify the effectiveness of the proposed algorithm. The experimental results show that higher fitting accuracy and generalization ability is found in the improved GPR whose hyperparameters are optimized by the Pelican optimization algorithm.

Prediction of product yields from lignocellulosic biomass pyrolysis based on gaussian process regression

The physicochemical characteristics of biomass, as well as the operational conditions of pyrolysis, are crucial factors that influence the production yield of tri-phase products during the pyrolysis of lignocellulosic biomass. Gaussian process regression (GPR) algorithms, being significant machine learning algorithms, are utilized for the prediction of pyrolytic product yields. These algorithms take into account the influence of biomass characteristics and pyrolysis conditions in a comprehensive manner. All of the prediction models exhibit strong performance, with R2 values exceeding 0.90 and RMSE values below 0.18. The study employed the Partial Dependence Plot (PDP) method to assess the influence patterns of individual features or interactions between two factors on the pyrolysis products. The findings indicate that the ultimate temperature reached during the pyrolysis process is a critical factor in influencing the generation of gaseous byproducts and solid residues. Specifically, higher temperatures are associated with increased production of gaseous byproducts and decreased production of solid residues. The optimal liquid phase yield was achieved at temperatures between 500 °C and 650 °C. In this research, the Particle Swarm Optimization (PSO) algorithm was employed to predict the optimal yields of pyrolysis products. This study offers a novel perspective on predicting product yields from the pyrolysis of lignocellulosic biomass with varying biomass characteristics and under different operational conditions.

Prediction of higher heating value of hydrochars using Bayesian optimization tuned Gaussian process regression based on biomass characteristics and process conditions

Hydrochars are valuable resources obtained from hydrothermal carbonization (HTC) of biomass. To optimize the reaction conditions of HTC, extensive experimentation is required, which is both costly and time consuming. In order to reduce the time and cost, this study develops new predictive models for higher heating values (HHV) of hydrochar based on Gaussian Process Regression (GPR), Ensemble, and Decision Tree (DT) algorithms using Bayesian Optimization (BO). This approach reduces prediction errors by combining GPR, Ensemble, and DT with BO. This is the first study on the application of BO for the hyperparameter selection as the basic learner. BO-GPR converged during training with the lowest Mean Absolute Error (0.1783) compared to BO-Ensemble (0.5128) and BO-DT (0.7430). The efficiencies of the algorithms were measured by the Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and coefficient of determination (R2). During the testing, BO-GPR generated MAE, RMSE and R2 of 0.4435, 0.5961 and 0.9705, respectively, and performed better than BO-Ensemble and BO-DT. The Nemenyi test showed that BO-GPR, BO-Ensemble, and BO-DT were statistically different in terms of their prediction ability. BO-GPR outperformed the other two methods.

Assessing the performance of the Gaussian Process Regression algorithm to fill gaps in the time-series of daily actual evapotranspiration of different crops in temperate and continental zones using ground and remotely sensed data

The knowledge of crop evapotranspiration is crucial for several hydrological processes, including those related to the management of agricultural water sources. In particular, the estimations of actual evapotranspiration fluxes within fields are essential to managing irrigation strategies to save water and preserve water resources. Among the indirect methods to estimate actual evapotranspiration, ETa, the eddy covariance (EC) method allows to acquire continuous measurement of latent heat flux (LE). However, the time series of EC measurements are sometimes characterized by a lack of data due to the sensors' malfunctions. At this aim, Machine Learning (ML) techniques could represent a powerful tool to fill possible gaps in the time series. In this paper, the ML technique was applied using the Gaussian Process Regression (GPR) algorithm to fill gaps in daily actual evapotranspiration. The technique was tested in six different plots, two in Italy, three in the United States of America, and one in Canada, with different crops and climatic conditions in order to consider the suitability of the ML model in various contexts. For each site, the climate variables were not the same, therefore, the performance of the method was investigated on the basis of the available information. Initially, a comparison of ground and reanalysis data, where both databases were available, and between two different satellite products, when both databases were available, have been conducted. Then, the GPR model was tested. The mean and the covariance functions were set by considering a database of climate variables, soil water status measurements, and remotely sensed vegetation indices. Then, five different combinations of variables were analyzed to verify the suitability of the ML approach when limited input data are available or when the weather variables are replaced with reanalysis data. Cross-validation was used to assess the performance of the procedure. The model performances were assessed based on the statistical indicators: Root Mean Square Error (RMSE), coefficient of determination (R2), Mean Absolute Error (MAE), regression coefficient (b), and Nash-Sutcliffe efficiency coefficient (NSE). The quite high Nash Sutcliffe Efficiency (NSE) coefficient, and the root mean square error (RMSE) low values confirm the suitability of the proposed algorithm.

Development and performance comparison of optimized machine learning-based regression models for predicting energy-related carbon dioxide emissions.

Accurate prediction of CO2 emissions for the countries has become a crucial task in decision-making processes for planning energy conversion and usage, supporting the design of effective emissions reduction strategies, and helping to achieve the goal of a sustainable and low-carbon future. Therefore, this study aims to develop a general model that can predict the national CO2 emissions of each country using data from 68 countries with high prediction accuracy based on machine learning regression models. Nine prediction models were developed using Support Vector Regression, Ensemble of Trees, and Gaussian Process Regression algorithms as machine learning methods, and their prediction performances were compared. Additionally, the hyperparameters of these three machine-learning methods were tuned by Bayesian optimization to improve their prediction performance. The test results of the optimized Gaussian Process Regression model (MSE = 106.68, RMSE = 10.328, MAE = 4.904, MAPE = 3.38%, R2 = 0.9998) showed that it was the best prediction model among the all developed models. Additionally, the optimized Gaussian Process Regression model gave very robust results in predicting CO2 emissions in many countries, indicating that it can be used reliably and with high accuracy as a promising prediction model.

Nonparametric Modeling and Control of Ship Steering Motion Based on Local Gaussian Process Regression

This paper aims to study the nonparametric modeling and control of ship steering motion. Firstly, the black box response model is derived based on the Nomoto model. Then, the establishment of a nonparametric response model and prediction of ship steering motion are realized by applying the local Gaussian process regression (LGPR) algorithm. To assess the performance of LGPR, two cases are studied, including a Mariner class vessel by using simulation data and a KVLCC2 tanker model by using experimental data. The results reveal that the response model identified by LGPR presents good prediction accuracy and low computational burden. Finally, the identified response model is used as the basis for developing the ship heading controller, and the results demonstrate that the proposed controller is able to achieve good dynamic performance.

Prediction of the thermal behavior of multi-walled carbon nanotubes-CuO-CeO2 (20-40-40)/water hybrid nanofluid using different types of regressors and evolutionary algorithms for designing the best artificial neural network modeling

For conducting an analysis of the experimental data, it is imperative to establish a mathematical correlation between the input and output variables. This entails executing a curve fitting or regression procedure on the data, for which numerous methodologies exist. Within the scope of present investigation, the design variables encompass the solid volume fraction (φ) and temperature. Thermal conductivity (TC) of MWCNT-CuO-CeO2 (20-40-40)/water hybrid nanofluid (HNF) is also the objective function. Ten different types of regressors are utilized for regression operations which are Multiple Linear Regression (MLR), Decision Tree (D-Tree), Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Extreme Learning Machine (ELM), Radial Basis Function (RBF), Adaptive Neuro-Fuzzy Inference System (ANFIS), Gaussian Process Regression (GPR), Multivariate Polynomial Regression (MPR) and Group Method of Data Handling (GMDH). Once the governing equations linking the design variables and the objective functions have been established, these equations can be employed to forecast the simulation data. By substituting the above input values into the equations, we can calculate the corresponding output values for the TC of the HNF. The results obtained from the MPR algorithm are compared to the experimental data. For the GPR, MLR, D-Tree, ELM, MPR, MLP, RBF, SVM, ANFIS, and GMDH algorithms, the maximum margin of error is found to be 0.031, 0.02579, 0.028946, 0.033889, 0.01568, 0.02515, 0.03485, 0.03, 0.0385, and 0.0178, respectively. Moreover, the kernel density estimation diagram indicates the gap between experimental data and data predicted by regression algorithms. Finally, it is evident that the MPR algorithm demonstrates to have a reduced residual dispersion, with the residuals approaching zero.

Operational optimization of copper flotation process based on the weighted Gaussian process regression and index-oriented adaptive differential evolution algorithm

Concentrate copper grade (CCG) is one of the important production indicators of copper flotation processes, and keeping the CCG at the set value is of great significance to the economic benefit of copper flotation industrial processes. This paper addresses the fluctuation problem of CCG through an operational optimization method. Firstly, a density-based affinity propagationalgorithm is proposed so that more ideal working condition categories can be obtained for the complex raw ore properties. Next, a Bayesian network (BN) is applied to explore the relationship between the operational variables and the CCG. Based on the analysis results of BN, a weighted Gaussian process regression model is constructed to predict the CCG that a higher prediction accuracy can be obtained. To ensure the predicted CCG is close to the set value with a smaller magnitude of the operation adjustments and a smaller uncertainty of the prediction results, an index-oriented adaptive differential evolution (IOADE) algorithm is proposed, and the convergence performance of IOADE is superior to the traditional differential evolution and adaptive differential evolution methods. Finally, the effectiveness and feasibility of the proposed methods are verified by the experiments on a copper flotation industrial process.

Adaptation of Deep Network in Transfer Learning for Estimating State of Health in Electric Vehicles during Operation

Battery state of health (SOH) is a significant metric for evaluating battery life and predicting battery safety. Currently, SOH research is largely based on laboratory data, with a dearth of research on electric vehicle (EV) operating data. Due to the difficulty in obtaining complete charge data under EV operating conditions, this study presents a SOH estimation method utilizing deep network adaptation. First, a data-driven approach is employed to extract voltage, current, state of charge (SOC), and incremental capacity (IC) data features. To compensate for the lack of aging information in the EV operation data domain, transfer learning is employed to construct the SOH estimation model. Additionally, to resolve inconsistent data distribution between the source laboratory battery data domain and the target EV operation data domain, an adaptive layer is added to the network, and adaptation of deep network (ADN) is utilized to enhance the model’s performance. Finally, the model is validated using electric bus operational data. Results indicate that this model’s average Mean Absolute Error (MAE) is less than 3.0%, and, compared to support vector machine (SVM) regression and Gaussian Process Regression (GPR) algorithms, the MAE is reduced by 27.7% and 38.4%, respectively.