Combination Of Support Vector Regression Research Articles

Predicting Rainfall-induced Landslide Using Bee Colony Algorithm Based on Support Vector Regression

Objective: Natural disasters caused by landslides have done great harm to agricultural production, people's lives, and property. Considering the slope disaster caused by heavy rainfall, it is important to establish an early warning system to monitor rainfall disaster prevention. Huafang University Slope Sustainable Development Research Center (HUSSDRC) has set up a meteorological station equipped with many sensors to provide early warning for landslides in Taiwan. Since the amount of data collected will soon become very large, there is a need to implement strong parallel frameworks containing information from the meteorological station and the displacement of tiltmeters required to predict the landslides caused by rainfall. Apache Spark (AS) is a general framework that contains the parallel process engine for data analytics. In this study, a hybrid method is utilized to predict rainfall-induced landslides. The proposed method combines support vector regression (SVR) with an artificial bee colony (ABC) algorithm on the parallel platform of AS. For the proposed method, the RMSE is 0.562, and it is the best value among these compared approaches. Methods: The SVR together with an ABC algorithm is applied to predict rainfall-induced landslides on AS. The AS can perform parallel data analytics in memory to speed up performance. However, it is hard to set up the best parameters for SVR. Thereafter, the ABC algorithm is utilized to search for the best parameters for SVR. Results: Compared with other methods, the proposed method results provide the smallest root mean square error (RMSE) for predicting rainfall-induced landslides. Conclusion: A hybrid method is proposed to predict rainfall-induced landslides. The proposed hybrid method is based on the parallel platform of AS in which SVR predicts the rainfall-induced landslides, and the ABC algorithm adjusts the best values of parameters for SVR. The comparison of RMSE for the method with existing approaches shows that the method indeed has the best value among compared approaches.

Flood mapping based on the combination of support vector regression and Heun’s scheme

In response to the increasing threats of flash flooding, how to generate flood maps in real time has drawn great attention. In this study, a hybrid ML model was developed for flood map generation with multiple lead times based on the combination of the support vector regression (SVR) method and Heun’s scheme for numerical correction, in which the hydraulic simulation results are used for model training while rainfall and flood sensor data are used as input variables (namely, the SVR-RFN model). To evaluate the performance of the hybrid model, two additional SVR models were developed and used as references for comparison—the first one used rainfall as input variables without sensor data (namely, the SVR-R model), and the second one added flood sensor data as input variables without numerical correction (namely, the SVR-RF model). The three models are tested and compared with a hydraulic model for a historical event in Sanyei River Basin, Taiwan. Using the hydraulic model as the benchmark, the hybrid model performed the best by reducing 63% of flood depth error and had the least deterioration of accuracy as lead time increased. The SVR-RF model performed the second by reducing the flood depth errors by 46% whereas the hydraulic model was more accurate than the SVR-R model. In a comparison of flood extents, the hydraulic model showed overestimation, the SVR-R model showed significant underestimation, whereas the two models with sensor data had balanced predictions. The flood maps generated by the hybrid model were applied for dike safety evaluation and showed a good capability in forecasting the risk of dike breach.

Calcium Soft Sensor Based on the Combination of Support Vector Regression and 1-D Digital Filter for Water Quality Monitoring

Calcium Soft Sensor Based on the Combination of Support Vector Regression and 1-D Digital Filter for Water Quality Monitoring

State of health estimation of power batteries based on multi-feature fusion models using stacking algorithm

The data-driven method is used widely to estimate the state of health (SOH) of the battery, but the selection of data features and the data training methods affect the estimation results greatly. With the stacking algorithm, this paper proposes a multi-feature fusion model to estimate battery SOH by fusing different feature parameters and combining support vector regression (SVR) and long short-term memory network (LSTM). The feature parameters were extracted only from the current change curve of the constant voltage charging stage. The support vector regression based on grid search (GS-SVR) was selected as the primary-learner, and the primary SVR models were constructed through 5-fold cross-validation for different feature parameters. The LSTM was selected as the secondary-learner. With the stacking algorithm, LSTM was used to fuse multiple primary SVR models to form an ensemble learner model to improve the performance of multi-feature fusion. The battery aging test data set and NASA battery test data set were used to evaluate the effectiveness. The results verified the validity and superiority of the proposed method. Compared with the existing estimation methods, root mean square error is reduced by at least 0.11, and mean absolute percentage error is reduced by at least 0.12%.

Prediction and feature analysis of punching shear strength of two-way reinforced concrete slabs using optimized machine learning algorithm and Shapley additive explanations

Punching shear strength (PSS) is an important index for the design and analysis of two-way reinforced concrete slabs. To accurately predict the PSS of two-way reinforced concrete slabs, a hybrid PSO-SVR model, which is the combination of support vector regression (SVR) and particle swarm optimization (PSO) algorithm was employed on 218 datasets with six design parameters as input variables and PSS as output. Moreover, the feature importance and the sensitivity analysis were performed to analyze quantitatively the feature importance and effect of the design parameters on the PSS. The results showed that, compared with the RMSE = 347.349, MAE = 210.019, and R 2 = 0.916 of the original SVR model, the PSO-SVR model reached better prediction performance with RMSE = 187.958, MAE = 135.889, and R 2 = 0.942. Among the six key design parameters, the effective depth of the slab D and the slab thickness H are the two main important factors that can cause large dispersion of the PSS in a stochastic environment and should be given more attention in the design and construction.

Remaining useful life prediction for aircraft engines based on phase space reconstruction and hybrid VNS-SVR model

Based on phase space reconstruction (PSR) and hybrid VNS-SVR model, a remaining useful life (RUL) prediction method for aircraft engines is proposed. The proposed hybrid model combines support vector regression (SVR), which has been successfully adopted for regression problems, with the variable neighborhood search (VNS). First, the phase space reconstruction is used to transform the selected one-dimensional performance sequences of aircraft engines into matrix forms, which increases the data information and improve the learning efficiency of the model effectively. Then, SVR is used to construct the prediction model. Meanwhile, a VNS algorithm is proposed to optimize the kernel parameters. Finally, the hybrid model is used to RUL prediction of the aircraft engines. The experimental results show that the method has a good prediction performance.

Application of variational mode decomposition and chaotic grey wolf optimizer with support vector regression for forecasting electric loads

Accurate electric load forecasting is critical in guaranteeing the efficiency of the load dispatch and supply by a power system, which prevents the wasting of electricity and facilitates energy sustainability. Applications of hybrid intelligent computing methods and swarm-based algorithms with the support vector regression (SVR) model are very promising for solving the problem of premature convergence. This paper presents a novel SVR-based electric load forecasting model by hybridizing variational mode decomposition (VMD), the chaotic mapping mechanism, and the grey wolf optimizer (GWO) in the VMD-SVR-CGWO model to improve the solution searching experiences and to determine the appropriate combination of SVR’s parameters that improve forecasting accuracy. Numerical examples that involve two widely known electric load data sets reveal that the proposed VMD-SVR-CGWO model outperforms other models with respect to forecasting accuracy.

Prediction of the axial compressive strength of circular concrete-filled steel tube columns using sine cosine algorithm-support vector regression

In this study, support vector regression optimized by a sine cosine algorithm was used to predict the axial bearing capacities of circular concrete-infilled steel tubular columns. A database consisting of 478 test specimens was constructed to train and validate the proposed model, and the effects of the data composition and the sine cosine optimization were investigated. The accuracy of the proposed method was verified by comparing the proposed model’s performance to those of other widely used machine learning methods and design formulae. The stability of the proposed method was verified by evaluating the average model performance of 100 random numerical experiments. For computer aided intelligent design, a backward prediction analysis using machine learning was performed, and the results indicate that the combination of support vector regression and the sine cosine algorithm can provide preliminary design assistance, whereas the empirical formulae cannot.

Intelligent control for energy conservation of air conditioning system in manufacturing systems

Energy conservation is a critical issue, particularly in manufacturing industries with intensive energy usage. The proportional–integral–derivative (PID) controller fails to adjust shop floor temperature against indoor heat dynamics which substantially affect the stability and robustness of the system to achieve energy conservation. By jointly considering the make-up unit (MAU) and a set of collaborative dry cooling coil (DCC) units of a large scaled industrial air condition system, an intelligent control model is proposed in the study to achieve energy conservation and ensure stability of shop floor temperature simultaneously. The temperature set point of the MAU, and the openness of massive distributed DCCs in the factory are optimized in real-time to guarantee the comfort of shop floor temperature, and minimize energy usage and damping. A combined support vector regression with genetic algorithm is proposed to solve the problem efficiently. In addition, the energy efficiency and robustness of the proposed optimization model in securing the stability of the shop floor temperature against indoor heat dynamics is proven through a case study. The total energy consumed by the proposed model is reduced by 7.09% compared with PID controller. The model can regulate shop floor temperature to facilitate operator comfort, product quality, and safe working environment.

An automatic real-time bus schedule redesign method based on bus arrival time prediction

Bus travel time often fluctuates greatly due to disturbance factors such as traffic conditions and passenger flow, which causes a reduction in the transit service as it cannot be provided as planned. Bus dispatching is necessary to handle this problem. Two crucial factors for real-time bus dispatching have been analysed in this study namely the bus arrival time prediction and timetable redesign. First, a bus arrival time prediction model combining Support Vector Regression (SVR) and Kalman Filter (K-SVR) was proposed. The headway is selected as an intermediate variable to improve the prediction accuracy by limiting the disturbance factors between two adjacent bus stops. Besides, by adjusting the parameters of the Kalman Filter, the predictive results are more suitable for practice to avoid the frequency of unnecessary timetable adjustment. Furthermore, an automatic timetable redesign method is given based on the proposed circle search algorithm, which can minimise the impact on the initial schedule by finding the minimum adjustment range in consideration of the normal working order. A case study in Shenzhen, China was conducted and the results verify that the K-SVR method could improve the prediction accuracy especially in peak hours. On this basis, in comparison with the traditional method, the timetable redesign algorithm could allocate the transit resources appropriately and perform well on indicators such as the headway fluctuation, compliance rate of headways, and schedule completeness. Moreover, it is more operable and valuable as a result of sending the departure instructions appropriately and maintaining the normal working order of drivers realistically.

A deep learning approach for prediction of air quality index in a metropolitan city

In India, the Central and State Pollution Control Boards have commissioned the National Air Monitoring Program (NAMP) which covers 240 cities with 342 monitoring stations. Air Quality Index (AQI) has been categorized into different groups. To predict the AQI in Chennai city, the Dataset was collected, then preprocessed to replace missing values and remove redundant data. The mean, mean square error and standard deviation are extracted using the Grey Level Co-occurrence Matrix (GLCM). The combination of Support Vector Regression (SVR) and Long Short-Term Memory (LSTM) based deep learning model is used to classify the AQI values. The proposed deep learning model gives an accurate and specific value for AQI on the city’s specified location compared to the existing techniques. The prediction accuracy is improved in the proposed deep learning method, which will caution the public to reduce to an acceptable level. The deep learning mechanism predicts the AQI values accurately and helps to plan the metropolitan city for sustainable development. The expected AQI value can control the pollution level by incorporating road traffic signal coordination, encouraging the people to use public transportation, and planting more trees on some locations.

A new hybrid algorithm model for prediction of internal corrosion rate of multiphase pipeline

Pipeline plays an important role in the oil and gas transportation industry. In recent years, more and more pipeline damages and breakdowns are caused by corrosion, which hurts the normal operation. Accurate prediction of the pipeline's corrosion rate is of great significance for the pipeline to operate safely and soundly. In this study, a hybrid intelligent algorithm method is proposed to predict the corrosion rate of the multiphase flow pipeline. The proposed model combines support vector regression (SVR), principal component analysis (PCA), and chaos particle swarm optimization (CPSO), named PCA-CPSO-SVR. PCA can reduce the data dimension and screen out the main variables of corrosion influencing factors. CPSO is utilized to optimize the hyperfine parameters in SVR, thus improving the prediction accuracy of the prediction model. The mean absolute error of the proposed model is 0.083, which is 18.6% lower than that of SVR. Compared with five benchmark models including linear regression (LR), artificial neural network (ANN), PCA-genetic algorithm-SVR, PCA-PSO-SVR, and De warred95(OLGA), the proposed model has higher prediction accuracy. According to the above results, PCA-CPSO-SVR has a good performance in the prediction of the corrosion rate of the multiphase flow pipeline.

Actuation response model from sparse data for wall turbulence drag reduction

Drag reduction of an actuated turbulent boundary layer at a momentum-thickness-based Reynolds number Re${}_{\ensuremath{\theta}}$ = 1000 is computed, modeled, and predicted. The drag reduction for the set of actuation parameters is modeled using 71 large-eddy simulations. This drag model allows extrapolation outside the actuation domain for larger wavelengths and amplitudes. The modeling novelty combines support vector regression for interpolation, a parametrized ridgeline leading out of the data domain, a scaling for the drag reduction, and a discovered self-similar structure of the actuation effect.

A support vector regression model hybridized with chaotic krill herd algorithm and empirical mode decomposition for regression task

This work presents a hybrid model that combines support vector regression (SVR), empirical mode decomposition (EMD), the krill herd (KH) algorithm and a chaotic mapping function. EMD is used to decompose input time series data into components with several intrinsic mode functions (IMFs) and one residual, to capture the trends in the input data. SVR is used to forecast separately IMFs and the residual owing to its effectiveness in solving nonlinear regression and time series problems. The KH algorithm is used to select the parameters in the SVR models. The Tent chaotic mapping function is hybridized with the KH algorithm to prevent premature convergence and increase the accuracy of the whole model. Two real-world datasets from the New South Wales (NSW, Australia) market and the New York Independent System Operator (NYISO, USA) are used to demonstrate the performance of the proposed EMD-SVRCKH model. The experimental results reveal that the proposed model provides competitive advantages over other models and offers greater forecasting accuracy.

Comparison of the Performance of Partial Least Squares and Support Vector Regressions for Predicting Fatty Acids and Fatty Acid Classes in Marine Oil Dietary Supplements by Using Vibrational Spectroscopic Data

Simple, fast, and accurate analytical techniques for verifying the accuracy of label declarations for marine oil dietary supplements containing eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) are required because of the increased consumption of these products. We recently developed broad-based partial least squares regression (PLS-R) models to quantify six fatty acids (FAs) and FA classes by using the spectroscopic data from a portable Fourier transform infrared (FTIR) device and a benchtop Fourier transform near infrared (FT-NIR) spectrometer. We developed an improved quantification method for these FAs and FA classes by incorporating a nonlinear calibration approach based on the machine learning technique support vector machines. For the two spectroscopic methods, high accuracy in prediction was indicated by low root mean square error of prediction and by correlation coefficients (R2) close to 1, indicating excellent model performance. The percent accuracy of the support vector regression (SV-R) model predicted values for EPA and DHA in the reference material was 90 to 110%. In comparison to PLS-R, SV-R accuracy for prediction of FA and FA class concentrations was up to 2.4 times higher for both ATR-FTIR and FT-NIR spectroscopic data. The SV-R models also provided closer agreement with the certified and reference values for the prediction of EPA and DHA in the reference standard. Based on our findings, the SV-R methods had superior accuracy and predictive quality for predicting the FA concentrations in marine oil dietary supplements. The combination of SV-R with ATR-FTIR and/or FT-NIR spectroscopic data can potentially be applied for the rapid screening of marine oil products to verify the accuracy of label declarations.

Hydrodynamic landslide displacement prediction using combined extreme learning machine and random search support vector regression model

Many models have been developed for landslide displacement prediction, but owing to complex landslide-formation mechanisms and landslide-inducing factors, such models have different prediction accuracies. Thus, landslide displacement prediction remains a popular but difficult topic of research. In this paper, a landslide prediction model is proposed by combining extreme learning machine (ELM) and random search support vector regression (RS-SVR) sub-models. Particularly, the combined model decomposed accumulative landslide displacement into two terms, trend and periodic displacements, using a time series model, and simulated and predicted the two terms using the ELM and RS-SVR sub-models, respectively. The predicted trend and periodic terms are then summed to obtain the total displacement. The ELM and RS-SVR sub-models are applied to predict the deformation of Baishuihe landslide in the Three Gorges Reservoir Area (TGRA) as an example. The results showed that the model effectively improved the accuracy, stability, and scope of application of landslide displacement prediction, thus providing a new method for landslide displacement prediction.

A combined support vector regression with firefly algorithm for prediction of bottom hole pressure

Bottom hole pressure (BHP) is a fundamental parameter for the proper design of the production process and the development of reservoirs. BHP can be measured directly through the deployment of pressure down-hole gauges (PDG) or by the application of existing correlations and mechanistic models based on surface measurements. Unfortunately, these methods suffer from two main problems: the cost of measurement which is quite expensive mainly for PDG, and the inaccuracies for the correlations and mechanistic models, due to the limitation of their ranges of application. In this work, a new model based on support vector regression (SVR) optimized with firefly algorithm (FFA) is proposed to predict BHP of vertical wells with multiphase flow. Firefly algorithm is implemented for the optimal selection of SVR hyper-parameters. SVR-FFA model development is done using real-life measurement datasets obtained from distinct Algerian oil wells. The performance of the SVR-FFA model is compared with another hybridization SVR-genetic algorithm, trial and error SVR and with existing correlations and mechanistic models. The results demonstrate that the SVR-FFA model outperforms all the other models.

A new model for predicting sulfur solubility in sour gases based on hybrid intelligent algorithm

The accurate prediction of solubility of elemental sulfur in high H2S-content gas (sour gas) is of critical importance in the exploitation of sour gas. Due to the measurement difficulties, high pressure and low sulfur content in the gas phase, there are limited experimental data about the solubility of sulfur in the literature to date. In order to determine the reliability of data about sulfur solubility in H2S, CO2 and CH4, an assessment test of experimental data for gas-solid system under high pressure is carried out. The assessment test is based on Gibbs-Duhem equation and P-R equation of state is used for modeling. The correlated parameters in the model are obtained by using chaos-based firefly algorithm (CFA). For the whole experimental data, the assessment results show that 28% data points are considered as thermodynamically consistent, 28% are inconsistent and 44% are deemed to be not fully consistent. After eliminating the unreliable data points, four optimization algorithms combined BP neural network and support vector regression (SVR) into eight hybrid intellect algorithms. The results show that the most accurate results can be obtained using CFA algorithm combined with support vector regression among eight hybrid intellect algorithms. Simultaneously, this new model can obtain more accurate results compared with previous proposed three empirical models. For sulfur solubility in sour gas, the result shows that the average relative deviation between the experimental data and calculated results (ARD) is 4.51%. For sulfur solubility in pure H2S and CO2, the ARDs are 2.11% and 10.12%, respectively.

Multigranular Wavelet Decomposition-Based Support Vector Regression and Moving Average Method for Service-Time Prediction on Web Map Service Platforms

The quality-of-service for web mapping depends on accurate service-time predictions when balancing cloud computing resources to ensure improved allocation, but the service-time on a web map service platform is nonstationary, periodic, and random, thereby making the classic prediction methods ineffective. To address this problem, we analyze the positive correlation between the service-time and access loads based on queuing theory, and build a multigranular service-time series according to variations in the volume of information contained at different time granularities. Furthermore, we propose a new wavelet decomposition (WD) based support vector regression (SVR) and moving average (MA) (WD-SVR-MA) model based on service-time decomposition theory. In order to predict the service-time accurately using our method, subsequences with nonstationary long-term trends and subsequences with stationary random fluctuations are extracted from the service-time series by WD, and predicted using the combined SVR and MA model. We compared the performance of the proposed method with other six models in experiments. Thick granularity and thin granularity service-time series were extracted from the log files of the “Tianditu” web map service platform. The proposed multigranular WD-SVR-MA model provides a variety of multigranular prediction results, demonstrating the superior performance and robustness, which are suitable for different scenarios.

Hybrid chemometric approach for estimating the heat of detonation of aromatic energetic compounds

This work presents an elegant technique for estimating the heat of detonation (HD) of thirty organic energetic compounds by combining support vector regression (SVR) and gravitational search algorithm (GSA). The work shows that numbers of nitrogen and oxygen atoms as well as the compound molar mass are sufficient as descriptors. On the basis of three performance measuring parameters, the hybrid GSA-SVR outperforms Mortimer and Kamlet (1968), Mohammad and Hamid (2004) and Mohammad (2006) models with performance improvement of 93.951%, 86.197%, and 47.104%, respectively. The superior performance demonstrated by the proposed method would be of immense significance in containing the potential damage of the explosives through quick estimation of HD of organic energetic compounds without loss of experimental precision.