Grid Search Support Vector Machine Research Articles

Extracting soil salinization information with a fractional-order filtering algorithm and grid-search support vector machine (GS-SVM) model

ABSTRACTThe remote sensing information on the extraction method is of great importance to improve the accuracy and efficiency of soil salinization information. The objective of this study is to develop remote sensing extraction techniques to improve soil salinization maps. The following procedures were used in this study: (1) developed a fractional-order algorithm-based methodology of filter from high-resolution remote sensing imagery (Sentinel-2 MSI); (2) investigated the changing trend of image under different order filters; and (3) used a grid-search algorithm-support vector machines (GS-SVM) classification to employ extraction information of soil salinization. The results showed that the Fractional-order filter method outperformed the integer derivative in extracted information of soil salinization. In comparison of the classification accuracy between fractional-order processing algorithm and integer-order image processing algorithm, the fractional order has improved remarkably. The optimal classification model was 0.6 order, 0.8 order, 1.4 order, 1.6 order, and 1.8 order models. The overall accuracy and kappa coefficient (κ) of these models are 91.90% and 0.90, respectively. Analysing and comparing between soil salt index and filtering algorithm (1.2 order), the researchers found that the classification results of the two methods are similar. In general, this method can successfully extract soil salinization information in dry regions.

An intelligent SVM modeling process for crude oil properties prediction based on a hybrid GA-PSO method

Abstract Properties prediction of crude oil remains an essential issue for refineries. In this communication, an exhaustive and extendable support vector machine (SVM) intelligent prediction process has been proposed to solve this problem. A novel hybrid genetic algorithm-particle swarm optimization (GA-PSO) method was applied to optimize the SVM model. The optimization process and result demonstrated that the newly proposed GA-PSO-SVM method was more accurate and time-saving than the classical GA or PSO method. Compared with the classical Grid-search SVM, the combined GA-PSO-SVM model appeared to be more applicable for the properties prediction task. The TBP distillation curve fitting was exampled to evaluate the performance of the developed model. The regression result demonstrated the high accuracy and efficiency of the proposed process. The model can be applied in the Industrial Internet as a plug-in, and the adaptability and flexibility is demonstrated by the implement of crude oil molecular reconstruction employing the intelligent prediction process.

A reliable method for colorectal cancer prediction based on feature selection and support vector machine.

Colorectal cancer (CRC) is a common cancer responsible for approximately 600,000 deaths per year worldwide. Thus, it is very important to find the related factors and detect the cancer accurately. However, timely and accurate prediction of the disease is challenging. In this study, we build an integrated model based on logistic regression (LR) and support vector machine (SVM) to classify the CRC into cancer and normal samples. From various factors, human location, age, gender, BMI, and cancer tumor type, tumor grade, and DNA, of the cancer, we select the most significant factors (p < 0.05) using logistic regression as main features, and with these features, a grid-search SVM model is designed using different kernel types (Linear, radial basis function (RBF), Sigmoid, and Polynomial). The result of the logistic regression indicates that the Firmicutes (AUC 0.918), Bacteroidetes (AUC 0.856), body mass index (BMI) (AUC 0.777), and age (AUC 0.710) and their combined factors (AUC 0.942) are effective for CRC detection. And the best kernel type is RBF, which achieves an accuracy of 90.1% when k = 5, and 91.2% when k = 10. This study provides a new method for colorectal cancer prediction based on independent risky factors. Graphical abstract Flow chart depicting the method adopted in the study. LR (logistic regression) and ROC curve are used to select independent features as input of SVM. SVM kernel selection aims to find the best kernel function for classification by comparing Linear, RBF, Sigmoid, and Polynomial kernel types of SVM, and the result shows the best kernel is RBF. Classification performance of LR + RF, LR + NB, LR + KNN, and LR + ANNs models are compared with LR + SVM. After these steps, the cancer and healthy individuals can be classified, and the best model is selected.

Optimization of support vector machine parameters for voltage stability margin assessment in the deregulated power system

Recently, the electric power systems are operated relatively close to their operational limits due to worldwide deregulated electricity market policies. The power systems are being operated with high stress, and hence sufficient voltage stability margin is necessary to be managed to ensure secure operation of the power system. A particle swarm optimization-based support vector machine (SVM) approach for online monitoring of voltage stability has been proposed in this paper. The conventional methods for voltage stability monitoring are less accurate and highly time-consuming consequently, infeasible for online application. SVM is a powerful machine learning technique and widely used in power system to predict the voltage stability margin, but its performances depend on the selection of parameters greatly. So, the particle swarm optimization is applied to determine the parameter settings of SVM. The proposed approach uses bus voltage angle and reactive power load as the input vectors to SVM, and the output vector is the voltage stability margin index. The effectiveness of the proposed approach is tested using the IEEE 14-bus test system, IEEE 30-bus test system and the IEEE 118-bus test system. The results of the proposed PSO-SVM approach for voltage stability monitoring are compared with artificial neural networks and grid search SVM approach with same data set to prove its superiority.

Feature Fusion of ICP-AES, UV-Vis and FT-MIR for Origin Traceability of Boletus edulis Mushrooms in Combination with Chemometrics.

Origin traceability is an important step to control the nutritional and pharmacological quality of food products. Boletus edulis mushroom is a well-known food resource in the world. Its nutritional and medicinal properties are drastically varied depending on geographical origins. In this study, three sensor systems (inductively coupled plasma atomic emission spectrophotometer (ICP-AES), ultraviolet-visible (UV-Vis) and Fourier transform mid-infrared spectroscopy (FT-MIR)) were applied for the origin traceability of 184 mushroom samples (caps and stipes) in combination with chemometrics. The difference between cap and stipe was clearly illustrated based on a single sensor technique, respectively. Feature variables from three instruments were used for origin traceability. Two supervised classification methods, partial least square discriminant analysis (FLS-DA) and grid search support vector machine (GS-SVM), were applied to develop mathematical models. Two steps (internal cross-validation and external prediction for unknown samples) were used to evaluate the performance of a classification model. The result is satisfactory with high accuracies ranging from 90.625% to 100%. These models also have an excellent generalization ability with the optimal parameters. Based on the combination of three sensory systems, our study provides a multi-sensory and comprehensive origin traceability of B. edulis mushrooms.

Storage Aided System Property Enhancing and Hybrid Robust Smoothing for Large-Scale PV Systems

This paper presents an energy scheduling and output smoothing scheme for storage aided utility scale photovoltaic systems. A weighted energy scheduling approach is adopted for the peak load periods, and this ensures enhanced performance with well-fitted supply-demand curve and flat net load variation. A novel smoothing method is proposed by blending double grid search support vector machine power prediction with first-in-first-out robust smoothing. The actual hourly and minute interval data sets for Australia are used for case studies, demonstrating the effectiveness and efficiency of the proposed scheme.

Dynamic measurement errors prediction for sensors based on firefly algorithm optimize support vector machine

The accuracy of dynamic measurement error predicting has the great significant influence on the precision and stability of the sensors. In order to solve the problem of the low accuracy of the model in traditional dynamic measurement error prediction, support vector machine (SVM) is applied to predicting the dynamic measurement error of sensors. However, in prediction tasks, a proper set of design parameters are essential for the performance and efficiency of SVM model. Hence, parameters selection for SVM by firefly algorithm (FA) is proposed in this study to avoid the local minimum value which occurs in the traditional method of parameter optimization. Two sensors of dynamic measurement error data are considered for modeling. Root mean squared error (RMSE) and mean absolute percentage error (MAPE) are employed to evaluate the performances of the models. These results are also compared with the results obtained from the grid search SVM (GS-SVM)and the particle swarm optimization SVM (PSO-SVM). Experimental results show that the model of SVM based on FA algorithm predicts more accurately and effectively on dynamic measurement errors prediction for sensors than the other models.

Improving the Diagnosis Accuracy of Hydrothermal Aging Degree of V2O5/WO3–TiO2 Catalyst in SCR Control System Using an GS–PSO–SVM Algorithm

Selective catalytic reduction (SCR) is one of the most effective technologies used for eliminating NOx from diesel engines. This paper presents a novel method based on a support vector machine (SVM) and particle swarm optimization (PSO) with grid search (GS) to diagnose the degree of aging of the V2O5/WO3–TiO2 catalyst in the SCR system. This study shows the aging effect on the performance of a NH3 slip based closed-loop SCR control system under different aging factors (α), which are defined by the SCR reaction rate ( R scr ). A diagnosis of the performance of GS–PSO–SVM has been presented as compared to SVM, GS–SVM and PSO–SVM to get reliable results. The results show that the average prediction diagnosis accuracy of the degree of catalytic aging is up to 93.8%, 93.1%, 92.9% and 92.0% for GS–PSO–SVM, PSO–SVM, GS–SVM and SVM respectively. It is demonstrated that GS–PSO–SVM is able to identify the SCR catalyst’s degree of aging, to ultimately assist with fault tolerance in the aging of the SCR catalyst.

Tool condition monitoring system based on support vector machine and differential evolution optimization

A tool condition monitoring system based on support vector machine and differential evolution is proposed in this article. In this system, support vector machine is used to realize the mapping between the extracted features and the tool wear states. At the same time, two important parameters of the support vector machine which are called penalty parameter C and kernel parameter [Formula: see text] are optimized simultaneously based on differential evolution algorithm. In order to verify the effectiveness of the proposed system, a multi-tooth milling experiment of titanium alloy was carried out. Cutting force signals related to different tool wear states were collected, and several time domain and frequency domain features were extracted to depict the dynamic characteristics of the milling process. Based on the extracted features, the differential evolution-support vector machine classifier is constructed to realize the tool wear classification. Moreover, to make a comparison, empirical selection method and four kinds of grid search algorithms are also used to select the support vector machine parameters. At the same time, cross validation is utilized to improve the robustness of the classifier evaluation. The results of analysis and comparisons show that the classification accuracy of differential evolution-support vector machine is higher than empirical selection-support vector machine. Moreover, the time consumption of differential evolution-support vector machine classifier is 5 to 12 times less than that of grid search-support vector machine.

Pattern Recognition of Finger Joint Angle for Intelligent Bionic Hand Using sEMG

Finger joint angle pattern recognition is significant for the development of an intelligent bionic hand. It makes the intelligent prosthesis understand the users intension more accurately and complete movements better. Surface electromyography signals have been widely used in intelligent bionics prosthesis research and rehabilitation medicine due to its advantages like high efficiency, convenient collection and non-invasive access. An improved grid-search method using a support vector machine has been proposed for the finger joint angle pattern recognition issue in surface electromyography signals. Pattern recognition for surface electromyography signals of index finger movement and metacarpophalangeal joint angle has been performed. Better classification performance was achieved through screening of feature vector combined with an improved grid-search support vector machine classification algorithm.

A novel quantitative structure–activity relationship method to predict the affinities of MT3 melatonin binding site

The linear regression (LR) and non-linear regression methods – grid search-support vector machine (GS-SVM) and projection pursuit regression (PPR) were used to develop quantitative structure–activity relationship (QSAR) models for a series of derivatives of naphthalene, benzofurane and indole with respect to their affinities to MT3/quinone reductase 2 (QR2) melatonin binding site. Five molecular descriptors selected by genetic algorithm (GA) were used as the input variables for the LR model and two non-linear regression approaches. Comparison of the results of the three methods indicated that PPR was the most accurate approach in predicting the affinities of the MT3/QR2 melatonin binding site. This confirmed the capability of PPR for the prediction of the binding affinities of compounds. Moreover, it should facilitate the design and development of new selective MT3/QR2 ligands.

Classification structure–activity relationship (CSAR) studies for prediction of genotoxicity of thiophene derivatives

The grid search support vector machine (GS-SVM) was used to build a classification structure–activity relationship (CSAR) model and to predict the genotoxicity property of 140 thiophene derivatives with the information derived from the compounds’ molecular structures. The seven descriptors selected by linear discriminant analysis (LDA) were used as the inputs to develop the GS-SVM model. Using the Grid Search method, a satisfactory model with a good predictive capability was obtained. The quality of the models was evaluated by the number of right classified compounds. The total accuracy of the LDA model was 81.4% and 85.2% for the training set and test set, respectively, and to the GS-SVM model was 92.9% and 92.6%, respectively. It was proved that the GS-SVM method was a very useful modeling approach with good classification ability for the genotoxicity of the thiophene derivatives. This work also provides a new idea and an alternative method to investigate the genotoxicity of the similar structures with thiophene derivatives, and can be extended to other toxicity studies.

Using classification structure pharmacokinetic relationship (SCPR) method to predict drug bioavailability based on grid-search support vector machine

The linear discriminant analysis (LDA) and the grid-search support vector machine (GS-SVM) were used to develop classification structure pharmacokinetic relationship models for predicting drug bioavailability. Bioavailability data for 167 compounds were taken from the literature, and the molecular descriptors were generated from the software CODESSA solely from molecular structures. Five descriptors selected by LDA were used to build the linear and nonlinear models. The obtained results confirmed the discriminative capacity of the calculated descriptors and the relationship with the drug bioavailability. The result of GS-SVM (total accuracy of 85.6%) was better than that of LDA (total accuracy of 72.4%), which indicated that the GS-SVM model was more reliable in the recognition of the drug bioavailability. The proposed method was very useful for the selection of new drugs products, and can also be extended in other classification structure pharmacokinetic relationship (CSPR) and classification structure activity relationship (CSAR) investigation.