Support Vector Machine Regression Model Research Articles

Evaluation of lipid oxidation of Chinese-style sausage during processing and storage based on electronic nose

A portable electronic nose was used for extracting flavour fingerprint map of Chinese-style sausage during processing and storage, in parallel with detection of acid value (AV) and peroxide value (POV) for evaluating lipid oxidation. Sausage samples during processing and storage were divided into three and five quality phases, respectively. After comparison of sensors response to lipid oxidation, optimal sensor array was determined. Several classification and regression models were developed to classify samples into their respective quality phase and predict lipid oxidation using full and optimal sensor array. Results indicated classification accuracy for sausage samples were, respectively, above 95% and 82% during the processing and storage. For support vector machine (SVM) and artificial neural networks (ANN) regression models, good performance in predicting AV and POV were obtained, with the coefficients of determination (R2s) >0.914 and 0.814 during processing and storage, respectively. Thus, E-nose demonstrated acceptable feasibility in evaluating the degree of lipid oxidation of Chinese-style sausage during processing and storage.

Retrieving Soybean Leaf Area Index from Unmanned Aerial Vehicle Hyperspectral Remote Sensing: Analysis of RF, ANN, and SVM Regression Models

Leaf area index (LAI) is an important indicator of plant growth and yield that can be monitored by remote sensing. Several models were constructed using datasets derived from SRS and STR sampling methods to determine the optimal model for soybean (multiple strains) LAI inversion for the whole crop growth period and a single growth period. Random forest (RF), artificial neural network (ANN), and support vector machine (SVM) regression models were compared with a partial least-squares regression (PLS) model. The RF model yielded the highest precision, accuracy, and stability with V-R2, SDR2, V-RMSE, and SDRMSE values of 0.741, 0.031, 0.106, and 0.005, respectively, over the whole growth period based on STR sampling. The ANN model had the highest precision, accuracy, and stability (0.452, 0.132, 0.086, and 0.009, respectively) over a single growth phase based on STR sampling. The precision, accuracy, and stability of the RF, ANN, and SVM models were improved by inclusion of STR sampling. The RF model is suitable for estimating LAI when sample plots and variation are relatively large (i.e., the whole growth period or more than one growth period). The ANN model is more appropriate for estimating LAI when sample plots and variation are relatively low (i.e., a single growth period).

NLOS identification and mitigation based on channel state information for indoor WiFi localisation

Indoor localisation could benefit greatly from non-line-of-sight (NLOS) identification and mitigation, since the major challenge for WiFi indoor ranging-based localisation technologies is multipath and NLOS. NLOS identification and mitigation on commodity WiFi devices, however, are challenges due to limited bandwidth and coarse multipath resolution with mere MAC layer received signal strength index. In this study, the authors explore and exploit the finer-grained PHY layer channel state information (CSI) to identify and mitigate NLOS. Key to the authors' approach is exploiting several statistical features of CSI, which are proved to be particularly effective. The approaches, NLOS identification support vector machine (NISVM) and related channel information regression model (RCIRM), based on machine learning are proposed to identify NLOS and mitigate NLOS error, respectively. Experiment results in various indoor scenarios with severe interferences demonstrated an overall NLOS identification rate of 94.12% with a false alarm rate of 5.88% and a better mitigation performance.

A Nondestructive Real-Time Detection Method of Total Viable Count in Pork by Hyperspectral Imaging Technique

A nondestructive method was developed for assessing total viable count (TVC) in pork during refrigerated storage by using hyperspectral imaging technique in this study. The hyperspectral images in the visible/near-infrared (VIS/NIR) region of 400–1100 nm were acquired for fifty pork samples, and their VIS/NIR diffuse reflectance spectra were extracted from the images. The reference values of TVC in pork samples were determined by classical microbiological plating method. Both partial least square regression (PLSR) model and support vector machine regression model (SVR) of TVC were built for comparative analysis to achieve better results. Different transformation methods and filtering methods were applied to improve the models. The results show that both the optimized PLSR model and SVR model can predict the TVC very well, while the SVR model based on second derivation was better, which achieved with RP (correlation coefficient of prediction set) = 0.94 and SEP (standard error of prediction set) = 0.4570 log CFU/g in the prediction set. An image processing algorithm was then developed to transfer the prediction model to every pixel of the image of the entire sample; the visualizing map of TVC would be displayed in real-time during the detection process due to the simplicity of the model. The results demonstrated that hyperspectral imaging is a potential reliable approach for non-destructive and real-time prediction of TVC in pork.

Surface Gloss Evaluation of Apples Based on Computer Vision and Support Vector Machine Method

This paper investigates computer vision applications for surface gloss evaluation to determine a quick surface gloss evaluation method for apples. “Red Fuji” apples were wax-coated with different concentrations of shellac solutions to obtain the apple samples with different levels of surface gloss. The surface gloss values and the color scales of the apple samples were detected using a pinhole gloss meter and a color meter. The apple sample images were captured and processed, and the color parameters of the high light areas were extracted. Support vector machine (SVM) regression and classification models were built to predict the surface gloss values and the surface gloss levels of apples, respectively. The results showed that to predict the surface gloss of apple samples, the correlation coefficients of the SVM regression model were 0.94 and 0.90 for the training and the testing groups, respectively. The classification accuracy rates of the SVM classification model for the training and the testing groups were 100 and 96.7%, respectively. Finally, apple surface gloss level classification software was developed, which showed good operating results for both classification accuracy rates and calculation speed. This paper provided a new surface gloss evaluation method based on computer vision for apples.

Automatic Facial Feature Extraction for Predicting Designers' Comfort With Engineering Equipment During Prototype Creation

Designers frequently utilize engineering equipment to create physical prototypes during the iterative concept generation and prototyping phases of design. Currently, evaluating designers' efficiency during prototype creation is a manual process that either involves observational or survey based approaches. Real-time feedback when using engineering equipment has the potential to enhance designers' efficiency or mitigate potential injuries that may result from incorrect use of equipment. Toward an automated approach to addressing these challenges, the authors of this work test the hypotheses that (i) there exists a difference in designers' comfort levels before and after they use a piece of engineering prototyping equipment and (ii) a machine learning model predicts the level of comfort a designer has while using engineering prototyping equipment with accuracies greater than random chance. It has been shown that the level of comfort that an individual has while completing a task impacts their performance. The authors investigate whether automatic tracking of designers' facial expressions during prototype creation predicts their level of comfort. A study, involving 37 participants using various engineering equipment, is used to validate the approach. The support vector machine (SVM) regression model yielded a range of R squared values from 0.82 to 0.86 for an equipment-specific model. A general model built to predict comfort level across all engineering equipment yielded an R squared value of 0.68. This work has the potential to transform the manner in which design teams utilize engineering equipment toward more efficient concept generation and prototype creation processes.

Study on Rail Profile Optimization Based on the Nonlinear Relationship between Profile and Wear Rate

This paper proposes a rail profile optimization method that takes account of wear rate within design cycle so as to minimize rail wear at the curve in heavy haul railway and extend the service life of rail. Taking rail wear rate as the object function, the vertical coordinate of rail profile at range optimization as independent variable, and the geometric characteristics and grinding depth of rail profile as constraint conditions, the support vector machine regression theory was used to fit the nonlinear relationship between rail profile and its wear rate. Then, the profile optimization model was built. Based on the optimization principle of genetic algorithm, the profile optimization model was solved to achieve the optimal rail profile. A multibody dynamics model was used to check the dynamic performance of carriage running on optimal rail profile. The result showed that the average relative error of support vector machine regression model remained less than 10% after a number of training processes. The dynamic performance of carriage running on optimized rail profile met the requirements on safety index and stability. The wear rate of optimized profile was lower than that of standard profile by 5.8%; the allowable carrying gross weight increased by 12.7%.

Empirical approach for strength prediction of geopolymer stabilized clayey soil using support vector machines

Potential of support vector machine regression (SVR) technique for strength estimation of geopolymer stabilized clayey soil has been investigated in the present paper. A comprehensive experimental database of 213 soil samples stabilized with ground granulated blast furnace slag (GGBS) based geopolymer binder were used to develop the SVR model. The database contains 28day unconfined compressive strength (UCS) results of soil samples generated with different combinations of experimental parameters. A comparative study of different kernel function on SVR model performance is discussed. The study showed that SVR is an effective tool for strength prediction of geopolymer stabilized clayey soil. Subsequently, a parametric study with SVR model was conducted to evaluate the effect of input parameters on UCS. Trends of the result obtained from parametric study were found to be in good agreement with previous research findings. Finally, using the SVR model, an empirical approach for strength prediction of GGBS based geopolymer stabilized clayey soil is proposed for practical application purpose.

Statistical methods and molecular docking for the prediction of thyroid hormone receptor subtype binding affinity and selectivity

The development of compounds that selectively modulate thyroid hormone action by serving as subtype-selective ligands of the thyroid hormone receptors (TRs) would be useful for clinical therapy. In the present work, quantitative structure-activity relationship (QSAR) models by adopting molecular descriptors to predict the TR binding activity were established based on a data set of TR ligands. The linear (multiple linear regression (MLR) and partial least squares regression (PLSR)) and nonlinear (support vector machine regression (SVR)) methods were employed to investigate the relationship between structural properties and binding activities. The proposed PLSR model was slightly superior to the MLR model and SVR model, as indicated by the reasonable statistical properties (TRβ: Rtr2 = 0.9594, Qte2 = 0.8091. TRα: Rtr2 = 0.9705, Qte2 = 0.8057). Additionally, molecular docking simulations were also performed to study the probable binding modes of the ligands and the TR subtype selectivity. The results indicate that substituents located in region A, region B, and region C and the orientation of these groups might result in the subtype selectivity based on the hydrogen bonding and electrostatic interactions. The derived QSAR models together with the molecular docking results have good potential in facilitating the discovery of novel TR ligands with improved activity and subtype selectivity.

A Personalized Electronic Movie Recommendation System Based on Support Vector Machine and Improved Particle Swarm Optimization.

With the rapid development of ICT and Web technologies, a large an amount of information is becoming available and this is producing, in some instances, a condition of information overload. Under these conditions, it is difficult for a person to locate and access useful information for making decisions. To address this problem, there are information filtering systems, such as the personalized recommendation system (PRS) considered in this paper, that assist a person in identifying possible products or services of interest based on his/her preferences. Among available approaches, collaborative Filtering (CF) is one of the most widely used recommendation techniques. However, CF has some limitations, e.g., the relatively simple similarity calculation, cold start problem, etc. In this context, this paper presents a new regression model based on the support vector machine (SVM) classification and an improved PSO (IPSO) for the development of an electronic movie PRS. In its implementation, a SVM classification model is first established to obtain a preliminary movie recommendation list based on which a SVM regression model is applied to predict movies’ ratings. The proposed PRS not only considers the movie’s content information but also integrates the users’ demographic and behavioral information to better capture the users’ interests and preferences. The efficiency of the proposed method is verified by a series of experiments based on the MovieLens benchmark data set.

A hybrid particle swarm optimization and support vector regression model for modelling permeability prediction of hydrocarbon reservoir

The significance of accurate permeability prediction cannot be over-emphasized in oil and gas reservoir characterization. Support vector machine regression (SVR), a computational intelligence technique, has been very successful in the estimation of permeability and has been widely deployed due to its unique features. However, careful selection of SVR hyper-parameters is highly essential to its optimum performance and this task is traditionally done using trial and error approach (TE-SVR) which takes a lot of time and do not guarantee optimal selection of the hyper-parameters. In this work, the performance of particle swarm optimization (PSO) technique, a heuristic optimization technique, is investigated for the optimal selection of SVR hyper-parameters for the first time in modelling and characterization of hydrocarbon reservoir. The technique is capable of automatic selection of the optimum combination of SVR hyper-parameters resulting in higher predictive accuracy and generalization ability of the developed model. The resulting PSO-SVR model is compared to SVR models whose parameters are obtained through random search (RAND-SVR) and trial and error approach (TE-SVR). The comparison is done using real-life industrial datasets obtained during petroleum exploration from four distinct oil wells located in a Middle Eastern oil and gas field. Simulation results indicate that the PSO-SVR model outperforms all the other models. Error reduction of 15.1%, 26.15%, 12.32% and 7.1% are recorded for PSO-SVR model compared to ordinary SVR (TE-SVR) in well-A, well-B, well-C and well-D, respectively. Also, reduction of 12.8%, 23.97%, 2.51% and 0.11 are recorded when PSO-SVR and RAND-SVR results are compared in the respective wells. Furthermore, the results show the potential of the application of heuristics algorithms, such as PSO, in the optimization of computational intelligence techniques employed in hydrocarbon reservoir characterizations. Therefore, PSO technique is proposed for the optimization of SVR hyper-parameters in permeability prediction and reservoir characterization based on its superior performance over the commonly employed optimization techniques.

Effects of Process Parameters on the Extraction of Quercetin and Rutin from the Stalks of Euonymus Alatus (Thumb.) Sieb and Predictive Model Based on Least Squares Support Vector Machine Optimized by an Improved Fruit Fly Optimization Algorithm

Ultrasonic-assisted extraction (UAE) of quercetin and rutin from the stalks of Euonymus alatus (Thunb.) Sieb in our laboratory, which aimed at evaluating and optimizing the process parameters, was investigated in this work. In addition, process parameters such as ethanol solution concentration, solvent volume/sample ratio, ultrasound power and extraction time, ultrasound frequency and extraction temperature were also first applied for evaluating the influence of extraction of quercetin and rutin. Optimum process parameters obtained were: ethanol solution 60%, extraction time 30 min, solvent volume/sample ratio 40 mL/g, ultrasound power 200 W, extraction temperature 30 °C and ultrasound frequency 80 kHz. Further a hybrid predictive model, which is based on least squares support vector machine (LS-SVM) in combination with improved fruit fly optimization algorithm (IFOA), was first used to predict the UAE process. The established IFOA-LS-SVM model, in which six process parameters and extraction yields of quercetin and rutin were used as input variables and output variables, respectively, successfully predicted the extraction yields of quercetin and rutin with a low error. Moreover, by comparison with SVM, LS-SVM and multiple regression models, IFOA-LS-SVM model has higher accuracy and faster convergence. Results proved that the proposed model is capable of predicting extraction yields of quercetin and rutin in UAE process.

Support Vector Regression Approach to Predict the Strength of FRP Confined Concrete

Support vector machine regression (SVR) technique was used for predicting the uniaxial compressive strength of FRP confined concrete. Two different SVR models, viz. SVR model for carbon fiber reinforced polymer confined concrete and SVR model for glass fiber reinforced polymer confined concrete, were developed. The prediction efficacy of developed SVR models was compared with that of ANN and existing empirical FRP strength prediction models. Furthermore, for assessing variable contributions in strength prediction process, a sensitivity analysis is also presented. The study showed that SVR could be a powerful alternative physical tool for strength prediction of FRP confined concrete.

BIQWS: efficient Wakeby modeling of natural scene statistics for blind image quality assessment

In this paper, a universal blind image quality assessment (IQA) algorithm is proposed that works in presence of various distortions. The proposed algorithm is a Blind Image Quality metric based on Wakeby Statistics (BIQWS) which extracts local mean subtraction and contrast normalization (MSCN) coefficients in spatial domain from input image. The MSCN coefficients are used for generating a Wakeby distribution statistical model to extract quality-aware features. The statistical studies indicate that the MSCN coefficients histogram is altered in the presence of various distortions with different severities. These changes are regular and can be used to estimate the type of the distortion and its severity. We extended our previous studies to extract efficient Wakeby distribution model parameters which are more sensitive to changes in MSCN coefficients. These parameters are used to form a quality-aware feature vector. This feature vector is then fed to an SVM (support vector machine) regression model with a nonlinear Kernel to predict the quality score of the input image without any information about the distortion type or reference image. Experimental results show that the image quality index obtained by the proposed method has higher correlation with respect to human perceptual opinions and it is superior in some distortions when compared to some full-reference and other state-of-the-art blind image quality assessment methods.

Prediction of Resilient Modulus of Lime-Treated Subgrade Soil Using Different Kernels of Support Vector Machine

AbstractThe resilient modulus (MR) plays a crucial role in mechanistic–empirical design such that acquiring the MR of lime-treated pavement layers seems to be necessary, because the use of lime materials in road projects is generally established. However, because of the complexity of and time and equipment requirements for repeated and cyclic load testing, several methods have been proposed to apply. In this paper, the novel artificial intelligence algorithm called support vector machine regression (SVR) has been applied to evaluate accurate values of lime-treated pavement layers’ MR. Moreover, polynomial kernel, radial basis function, and linear kernel as three different kernels of SVR were used to predict the MR of lime-treated subgrade soil. To create the model and validate the algorithm’s performance, approximately 75% of the data was selected as training data sets, and the remaining ones were applied as testing data sets. For this study, the obtained results indicate that developed SVR models produce...

Empirical Prediction of Leaf Area Index (LAI) of Endangered Tree Species in Intact and Fragmented Indigenous Forests Ecosystems Using WorldView-2 Data and Two Robust Machine Learning Algorithms

Leaf area index (LAI) is an important biophysical trait for forest ecosystem and ecological modeling, as it plays a key role for the forest productivity and structural characteristics. The ground-based methods like the handheld optical instruments for predicting LAI are subjective, pricy and time-consuming. The advent of very high spatial resolutions multispectral data and robust machine learning regression algorithms like support vector machines (SVM) and artificial neural networks (ANN) has provided an opportunity to estimate LAI at tree species level. The objective of the this study was therefore to test the utility of spectral vegetation indices (SVI) calculated from the multispectral WorldView-2 (WV-2) data in predicting LAI at tree species level using the SVM and ANN machine learning regression algorithms. We further tested whether there are significant differences between LAI of intact and fragmented (open) indigenous forest ecosystems at tree species level. The study shows that LAI at tree species level could accurately be estimated using the fragmented stratum data compared with the intact stratum data. Specifically, our study shows that the accurate LAI predictions were achieved for Hymenocardia ulmoides using the fragmented stratum data and SVM regression model based on a validation dataset (R2Val = 0.75, RMSEVal = 0.05 (1.37% of the mean)). Our study further showed that SVM regression approach achieved more accurate models for predicting the LAI of the six endangered tree species compared with ANN regression method. It is concluded that the successful application of the WV-2 data, SVM and ANN methods in predicting LAI of six endangered tree species in the Dukuduku indigenous forest could help in making informed decisions and policies regarding management, protection and conservation of these endangered tree species.

Parameter Identification of Ship Maneuvering Model Based on Support Vector Machines and Particle Swarm Optimization

Combined with the free-running model tests of KVLCC ship, the system identification (SI) based on support vector machines (SVM) is proposed for the prediction of ship maneuvering motion. The hydrodynamic derivatives in an Abkowitz model are determined by the Lagrangian factors and the support vectors in the SVM regression model. To obtain the optimized structural factors in SVM, particle swarm optimization (PSO) is incorporated into SVM. To diminish the drift of hydrodynamic derivatives after regression, a difference method is adopted to reconstruct the training samples before identification. The validity of the difference method is verified by correlation analysis. Based on the Abkowitz mathematical model, the simulation of ship maneuvering motion is conducted. Comparison between the predicted results and the test results demonstrates the validity of the proposed methods in this paper.

A Method for Missing Data Recovery of Air Pollutants Monitoring in Henhouse Based on QGSA-SVM

To solve the data missing problem caused by sensor faults during the air pollutants monitoring in henhouse, a method for missing data recovery was proposed based on support vector machine (SVM). Multiple factors that influence monitoring values of the air pollutants in henhouse, such as temporal, spatial and environmental, were considered to established a SVM regression model to estimate the missing data of the air pollutants monitoring. Meanwhile, to obtain better prediction accuracy, regression model parameters were optimized by a novel hybrid optimization algorithm which was combined standard genetic algorithm with quantum genetic strategy and simulated annealing tactics. Taking the data processing of the ammonia (NH3) concentration as an example, the proposed method was tested with the monitoring data of 3 days in a farm. The estimation results of missing data shown that there was a litter error between the estimated data and monitoring data, the maximal relative error was 5.87% (percent), the average relative error was 1.77% (percent). It is verified that this method of missing data recovery is feasible and valid.

Comparing different data preprocessing methods for monitoring soil heavy metals based on soil spectral features

The lands near mining industries in the Czech Republic are subjected to soil pollution with heavy metals. Excessive heavy metal concentrations in soils not only dramatically impact the soil quality, but also due to their persistent nature and indefinite biological half-lives, potentially toxic metals can accumulate in the food chain and can eventually endanger human health. Monitoring and spatial information of these elements require a large number of samples and cumbersome and time-consuming laboratory measurements. A faster method has been developed based on a multivariate calibration procedure using support vector machine regression (SVMR) with cross-validation, to establish a relationship between reflectance spectra in the visible-near infrared (Vis-NIR) region and concentration of Mn, Cu, Cd, Zn, and Pb in soil. Spectral preprocessing methods, first and second derivatives (FD and SD), standard normal variate (SNV), multiplicative scatter correction (MSC), and continuum removal (CR) were employed after smoothing with Savitzky-Golay to improve the robustness and performance of the calibration models. According to the criteria of maximal coefficient of determination (R2cv) and minimal root mean square error of prediction in cross-validation (RMSEPcv), the SVMR algorithm with FD preprocessing was determined as the best method for predicting Cu, Mn, Pb, and Zn concentration, whereas the SVMR model with CR preprocessing was chosen as the final method for predicting Cd. Overall, this study indicated that the Vis-NIR reflectance spectroscopy technique combined with a continuously enriched soil spectral library as well as a suitable preprocessing method could be a nondestructive alternative for monitoring of the soil environment. The future possibilities of multivariate calibration and preprocessing with real-time remote sensing data have to be explored.

Support vector machine to predict the indirect tensile strength of foamed bitumen-stabilised base course materials

Indirect tensile strength (ITS) crucially plays a significant role as a main feature for mix design of foamed bitumen-stabilised base course layers. Due to time-consuming and equipment requirements for conducting the ITS test, it appears to be demanded to develop the numerical models in order to predict the ITS of a foamed bitumen-stabilised base course. In this paper, the novel artificial intelligence algorithm of support vector machine regression (SVR) has been applied to predict the accurate value of indirect tensile strength (ITS) of the foamed bitumen-stabilised base course. Moreover, two kernels of SVR including the polynomial kernel and radial basis function (RBF) kernel have been investigated to predict the accurate ITS values of both unsoaked and soaked foamed bitumen-stabilised base separately. In order to create the model and to validate the algorithm performance, about 80% of data were randomly selected as the training data set and the remaining ones applied as the testing data set. The obtained results indicate that the developed SVR models produce a high prediction ability for this study. In addition, the obtained predicting correlation coefficients (R2) of polynomial and RBF kernels for both unsoaked and soaked samples were compared individually and, eventually, the RBF kernel of the SVR model was selected as the most accurate computational predicting model.