Improved Support Vector Regression Model Research Articles

Groundwater level prediction using an improved SVR model integrated with hybrid particle swarm optimization and firefly algorithm

The demand for water resources has increased due to rapid increase of metropolitan areas brought on by growth in population and industrialisation. In addition, the groundwater recharge is being afftected by shifting land use pattern caused by urban development. Using precise and trustworthy estimates of groundwater level is vital for the sustainable groundwater resources management in the face of changing climatic circumstances. In this context, machine learning (ML) methods offer a new and promising approach for accurately forecasting long-term changes in the groundwater level (GWL) without computational effort of developing a comprehensive flow model. In order to simulate GWL, five data-driven (DD) models, including the hybridization of support vector regression (SVR) with two optimisation algorithms i.e., firefly algorithm and particle swarm optimisation (FFAPSO), SVR-FFA, SVR-PSO, SVR and Multilayer perception (MLP), have been examined in the present study. Spatial clustering was utilised to choose four observation wells within Cuttack district in order to study and assess the water levels. Six scenarios were created by incorporating numerous variables, such as GWL in the previous months, evapotranspiration, temperature, precipitation, and river discharge. The goal was to identify the variables that were most efficient in predicting GWL. The SVR-FFAPSO model performs best in GWL forecasting for Khuntuni station, according to the quantitative analysis with correlation coefficient (R) = 0.9978, Nash–Sutcliffe efficiency (NSE) = 0.9933, mean absolute error (MAE) = 0.00025 (m), root mean squared error (RMSE) = 0.00775 (m) during the training phase. It is advised that groundwater monitoring network and data collecting system are strengthen in India for ensuring effective modelling of long-term management of groundwater resources.

Research on Yield Prediction Technology for Aerospace Engine Production Lines Based on Convolutional Neural Networks-Improved Support Vector Regression

Improving the prediction accuracy of aerospace engine production line yields is of significant importance for enhancing production efficiency and optimizing production scheduling in enterprises. To address this, a novel method called Convolutional Neural Networks-Improved Support Vector Regression (CNN-ISVR) has been proposed for predicting the production line yield of aerospace engines. The method first divides the factors influencing production line yield into production cycle and real-time status information of the production line and then analyzes the key feature factors. To solve the problem of poor prediction performance in traditional SVR models due to the subjective selection of kernel function parameters, an improved SVR model is presented. This approach combines the elite strategy genetic algorithm with the hyperparameter optimization method based on grid search and cross-validation to obtain the best penalty factor and kernel function width of the Radial Basis Function (RBF) kernel function. The extracted features of production data are then used for prediction by inputting them into the improved support vector regression model, based on a shallow CNN without dimensionality reduction. Finally, the prediction accuracy of the CNN-ISVR model is evaluated using the three commonly used evaluation metrics: Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE) and coefficient of determination (R2). The model’s prediction results are then compared to those of other models. The CNN-ISVR hybrid model is shown to outperform other models in terms of prediction accuracy and generalization ability, demonstrating the effectiveness of the proposed model.

Estimation of Spring Maize Evapotranspiration in Semi-Arid Regions of Northeast China Using Machine Learning: An Improved SVR Model Based on PSO and RF Algorithms

Accurate estimation of crop evapotranspiration (ETc) is crucial for effective irrigation and water management. To achieve this, support vector regression (SVR) was applied to estimate the daily ETc of spring maize. Random forest (RF) as a data pre-processing technique was utilized to determine the optimal input variables for the SVR model. Particle swarm optimization (PSO) was employed to optimize the SVR model. This study used data obtained from field experiments conducted between 2017 and 2019, including crop coefficient and daily meteorological data. The performance of the innovative hybrid RF–SVR–PSO model was evaluated against a standalone SVR model, a back-propagation neural network (BPNN) model and a RF model, using different input meteorological variables. The ETc values were calculated using the Penman–Monteith equation, which is recommended by the FAO, and used as a reference for the models’ estimated values. The results showed that the hybrid RF–SVR–PSO model performed better than all three standalone models for ETc estimation of spring maize. The Nash–Sutcliffe efficiency coefficient (NSE), root mean square error (RMSE), mean absolute error (MAE) and coefficient of determination (R2) ranges were 0.956–0.958, 0.275–0.282 mm d−1, 0.221–0.231 mm d−1 and 0.957–0.961, respectively. It is proved that the hybrid RF–SVR–PSO model is appropriate for estimation of daily spring maize ETc in semi-arid regions.

Prediction Method for Sugarcane Syrup Brix Based on Improved Support Vector Regression

The brix of syrup is an important parameter in sugar production. To accurately measure syrup brix, a novel measurement method based on support vector regression (SVR) is presented. With the resonant frequency and quality factor as inputs and syrup brix as the output, a mathematical model of the relationship between the resonant frequency, quality factor, and syrup brix is established. Simultaneously, the particle swarm optimization (PSO) algorithm is used to optimize the penalty coefficient and radial basis kernel function of SVR to improve the performance of the model. The calculation model is trained and tested using the collected experimental data. The results show that the mean absolute error, mean absolute percentage error, and root mean square error of the syrup brix calculation model based on the improved SVR model can reach 0.74 °Bx, 2.24%, and 0.90 °Bx, respectively, while the determination coefficient can reach 0.9985. The simulation of the online measurement of syrup brix in the actual production process proves the excellent prediction performance of the syrup brix calculation model based on the improved PSO–SVR model, which can thus be used to predict the syrup brix.

Fatigue life prediction of welded joints based on improved support vector regression model under two‐level loading

AbstractTo meet the lightweight needs of modern machinery equipment, welded structure is widely used in practical engineering. During the service period, the welded structure is often subjected to multilevel variable amplitude loading, and fatigue failure is easy to occur. Due to the complexity and stochasticity of fatigue process, the traditional life prediction models cannot meet the demand of practical engineering. In this work, a novel prediction model based on the whale optimization algorithm (WOA) together with the support vector regression (SVR), namely, a WOA‐SVR model, is established. It combines the advantages of WOA and SVR for global searching optimization and high prediction accuracy dealing with small sample size. By taking into account the sequence of loadings and the interaction effect between loadings, the proposed WOA‐SVR model could overcome the deficiencies and shortcomings in existing models. Experiment results show that WOR‐SVR has better prediction performance than five models.

Self-discharge prediction method for lithium-ion batteries based on improved support vector machine

An improved support vector regression (SVR) method is proposed for predicting the self-discharge voltage drop (SDV-drop) in lithium-ion batteries. Multiple features were extracted according to the charge and discharge curves of lithium-ion batteries, and the three features having the strongest correlation with the SDV-drop were identified via grey relational analysis. Then, these three features were assigned different weight parameters to obtain composite features which were input into the improved support vector machine through differential evolution algorithm parameter optimization training. Finally, the improved SVR model was obtained. Model training and testing were performed via a battery charge and discharge experiment and battery static experimental data of a new energy vehicle company, and the results indicated that the proposed method had a higher prediction accuracy than the neural-network model and the Gaussian process regression model.

Icing Time Prediction Model of Pavement Based on an Improved SVR Model with Response Surface Approach

Pavement icing imposes a great threat to driving safety and impacts the efficiency of the road transportation system in cold regions. This has attracted research predicting pavement icing time to solve the problems brought about by icing. Different models have been proposed in the past decades to predict pavement icing, within which support vector regression (SVR) is a widely used algorithm for calibrating highly nonlinear relationships. This paper presents a hybrid improved SVR algorithm to predict the time of pavement icing with an enhancement operation by response surface method (RSM) and particle swarm optimization (PSO). RSM is used to increase the number of input data collected onsite. Based on that, the optimal SVR model is established by optimizing the kernel function parameters and penalty coefficient with the particle swarm optimization (PSO) algorithm. The hybrid improved SVR is compared with SVR, PSO-SVR, and RSM-PSO for coefficient of determination (R2), mean absolute error, mean absolute percentage error, and root mean square error to check the effectiveness of PSO and RSM in optimizing SVR. The results show that the combination of two methods in the hybrid improved algorithm has a better optimization capability with R2 of 0.9655 and 0.9318 in a train set and test set, respectively, which outperforms PSO-SVR, RSM-SVR, and SVR. In addition, the R2 of the hybrid improved SVR and PSO-SVR both reach the optimal fitness value approximately at the iteration of 20, which suggests that convergence capacity remains relatively constant with the predictive accuracy being improved.

Prediction of Welding Deformation and Residual Stress of a Thin Plate by Improved Support Vector Regression.

Thin plates are widely utilized in aircraft, shipbuilding, and automotive industries to meet the requirements of lightweight components. Especially in modern shipbuilding, the thin plate structures not only meet the economic requirements of shipbuilding but also meet the strength and rigidity requirements of the ship. However, a thin plate is less stable and prone to destabilizing deformation in the welding process, which seriously affects the accuracy and performance of the thin plate welding structure. Therefore, it is beneficial to predict welding deformation and residual stress before welding. In this paper, a thin plate welding deformation and residual stress prediction model based on particle swarm optimization (PSO) and grid search(GS) improved support vector regression (PSO-GS-SVR) is established. The welding speed, welding current, welding voltage, and plate thickness are taken as input parameters of the improved support vector regression model, while longitudinal and transverse deformation and residual stress are taken as corresponding outputs. To improve the prediction accuracy of the support vector regression model, particle swarm optimization and grid search are used to optimize the parameters. The results show that the improved support regression model can accurately evaluate the deformation and residual stress of butt welding and has important engineering guiding significance.

Prediction of tailgate stability in mechanized longwall mines using an improved support vector regression model

Tailgate stability in a mechanized longwall mine is serious to mine productivity and personnel’s safety. Although the stability of both roadways is critical in longwall mining, the tailgate is subjected to higher stresses and deformations than the other one. Therefore, the prediction of tailgate stability is a distinctive challenge in mechanized coal mining to ensure its functionality during mining operations. In this regard, an Improved Support Vector Regression (ISVR) model was developed to sustain and secure longwall mine design through stability prediction of the tailgate roadway based on the roof displacements and geomechanical data. For model development, the geomechanical information gathered through laboratory tests and site investigations in Tabas coal mine, Iran, was introduced to the ISVR model as independent variables. The roof displacements values, which were monitored in a 1.2 km long tailgate, were also used as the dependent variable. According to the results, the proposed ISVR model could predict roof displacements in a reasonable accordance with measured ones. The squared correlation coefficient (R2) between ISVR predicted and measured roof displacements showed a high conformity with R2 = 0.91. The results of the ISVR model were compared with those of Artificial Neural Networks (ANNs) and Multivariable Linear Regression (MLR), which respectively yielded R2 = 0.87 and R2 = 0.81. In conclusion, the ISVR model appears to be a proper measure for indicating unstable zones ahead of time in mechanized and high-speed longwall mining.

Prediction of household electricity consumption and effectiveness of concerted intervention strategies based on occupant behaviour and personality traits

Household electricity consumption influenced by various behavioural intervention strategies is difficult to predict due to the uncertainty that arises from human behaviours and their responses to intervention. Based on an energy conservation experiment conducted in Hangzhou, China, the study aims to develop an improved Support Vector Regression model that is capable of predicting household electricity consumption under multiple intervention strategies. This paper firstly proposes a variable selection approach to determine the best subset of consumption predictors using Akaike Information Criterion. The critical predictors of energy behaviours, personality traits, demographic/building features, weather indicators and the historical monthly consumption is identified in this process. Furthermore, this research also introduces the interaction effect between the energy behaviour and all other predictors to the Support Vector Regression model which applies Gaussian radial basis function optimised by genetic algorithm as the kernel function. More importantly, the proposed model is able to select the optimal intervention strategy and to predict the maximum electricity savings for each household. It enables the households to achieve an average reduction of 12.1% in monthly electricity consumption compared with the conventional behavioural intervention. Moreover, the Monte Carlo simulation is performed to explore the relationship between personality traits, the best-fit intervention strategies and the maximum electricity savings. The study also identifies five types of households with different combinations of extraversion and conscientiousness that respond differently to the optimised interventions. The findings of this study contribute to the residential demand-side energy management by enriching and diversifying personalised behavioural intervention strategies.

Reliability Optimization of Structural Deformation with Improved Support Vector Regression Model

Deformation is one important failure mode of turbine blades. The quality of a model seriously influences the reliability optimization of turbine blades in turbo machines. To improve the reliability optimization of turbine blades, this paper proposes a novel machine learning-based reliability optimization approach, named improved support vector regression (SR) model (ISRM) method, by fusing artificial bee colony (ABC), traditional SR model, and multipopulation genetic algorithm (MPGA). In this proposed method, the ABC algorithm was applied to find the optimal parameters in the SR model to establish accurate ISRM, following the thought of the surrogate model method and the randomness of constraints. Then the reliability optimization model and procedure with the ISRM method were resolved by the MPGA. Regarding many design parameters (i.e., rotor speed, temperature, and aerodynamic pressure), design objective (i.e., blade deformation), and the randomness constraints of reliability degree and boundary conditions, we performed the reliability optimization of a turbine blade deformation. From the optimization results, we find that the turbine blade deformation is reduced by 0.09329 mm, and the ISRM learning method can improve the reliability optimization design of complex structures with the emphasis on modeling precision and optimization efficiency. The works of this paper provide a machine learning-based reliability optimization approach for the reliability optimization of complex structures and enrich and develop mechanical reliability theory and methods.

Optimization of smooth blasting parameters for mountain tunnel construction with specified control indices based on a GA and ISVR coupling algorithm

The smooth blasting method has been widely used in the construction of mountain tunnels to decrease the volume of overbreak or underbreak and maintain the tunnel outline in the design shape. However, due to the shortcomings of existing optimization theories and the complexity of rock masses, optimizing the smooth blasting parameters in arbitrary geological conditions with specified control indices is challenging. Eighteen on-site smooth blasting experiments were conducted during the construction of the long Foling highway tunnel in China. These experimental data were used as the training samples for machine learning. By training these samples, an improved support vector regression (ISVR) model was proposed to map the relation between the inputs, comprising the geological conditions (the basic quality [BQ] grade of the rock mass, saturated uniaxial compression strength of rock, and overburden depth) and control indices and the outputs of the smooth blasting parameters, including the spacing of perimeter holes and relief holes, minimum burden and linear charge concentration of perimeter holes. A genetic algorithm (GA) was coupled with an ISVR algorithm to automatically search the optimal parameters of the ISVR model during the training process. Using the ISVR model, the optimization of smooth blasting parameters can be obtained based on certain geological conditions of surrounding rock and specified control indices, including the crown settlement, thickness of the blasting damage zone (BDZ) in which the travelling velocity of ultrasonic waves is reduced significantly due to explosive vibration, volume of overbreak or underbreak, and radial decoupling ratio. According to the application results of the Foling tunnel, the ISVR model was shown to be superior since it can outperform certain existing models. As geological conditions and control indices are comprehensively considered, the proposed ISVR model of smooth blasting parameters is expected to be more feasible and reliable and is thus recommended for use in similar tunnel projects.

Application of improved support vector regression model for prediction of deformation modulus of a rock mass

Deformation modulus of a rock mass is one of the crucial parameters used in the design of surface and underground rock engineering structures. Determination of this parameter by testing cylindrical core samples is almost impossible due to the presence of discontinuities. Due to the problems in determining the deformability of jointed rock masses at the laboratory-scale, various in situ test methods such as plate loading tests, dilatometer etc. have been developed. Although these methods are currently the best techniques, they are expensive and time-consuming, and present operational problems. To overcome this difficulty, in this paper, presents the results of the application of hybrid support vector regression (SVR) with harmony search algorithm , differential evolution algorithm and particle swarm optimization algorithm (PSO). The optimized models were applied to available data given in open source literature and the performance of optimization algorithm was assessed by virtue of statistical criteria. In these models, rock mass rating (RMR), depth, uniaxial compressive strength of intact rock (UCS) and elastic modulus of intact rock (Ei) were utilized as the input parameters, while the deformation modulus of a rock mass was the output parameter. The comparative results revealed that hybrid of PSO and SVR yield robust model which outperform other models in term of higher squared correlation coefficient (R2) and variance account for (VAF) and lower mean square error (MSE), root mean squared error (RMSE) and mean absolute percentage error (MAPE).

An improved support vector regression model for estimation of saturation pressure of crude oils

Use of intelligence based approach for modeling of crude oil saturation pressure is viable alternative since this parameter plays influential role in the reservoir calculation. The objective of current study is to develop a smart model based on fusing of support vector regression model and optimization technique for learn the relation between the saturation pressure and compositional data viz. temperature, hydrocarbon and non-hydrocarbon compositions of crudes, and heptane-plus specifications. The optimization methods improve performance of the support vector regression (SVR) model through finding the proper value of their free parameters. The optimization methods which embedded in the SVR formulation in this study are genetic algorithm (GA), imperialist competitive algorithm (ICA), particle swarm optimization algorithm (PSO), cuckoo search algorithm (CS), and bat-inspired algorithm (BA). The optimized models were applied to experimental data given in open source literatures and the performance of optimization algorithm was assessed by virtue of statistical criteria. This evaluation resulted clearly show the superiority of BA when integrated with support vector regression for determining the optimal value of its parameters. In addition, the results of aforementioned optimized models were compared with currently available predictive approaches. The comparative results revealed that hybrid of BA and SVR yield robust model which outperform other models in term of higher correlation coefficient and lower mean square error.

An improved support vector model in car-rental revenue forecast

Monthly revenue per unit (RPU), measured by the total monthly revenue divided by the total monthly number of cars in the fleet, is one of the most popular profitability assessments in the car-rental industry. Because pricing decision is influenced by profit a lot, accurate forecast of RPU is thus useful for pricing in car-rental industry. Support vector regression (SVR) is an emerging forecast technique and successfully applied in many domains. The aim of this study is to investigate the feasibility of SVR in RPU forecast. Because SVR model contains three parameters influencing the forecasting accuracy, this work treats the forecasting error as an objective and employs tabu search (TS) algorithms to minimize the forecasting error. Actual RPU data are used to depict the forecasting performance of the improved SVR model. Two forecasting model, namely the general regression neural network (GRNN) model and the autoregressive integrated moving average (ARIMA) model, are used to forecast the same RPU data. Simulation results indicate that the improved SVR model is superior to the other models in terms of forecasting accuracy.