Accuracy Of Support Vector Regression Research Articles

Gold Price Prediction Using Support Vector Regression

In this modern era, one of the businesses that continues to grow is investment. Gold has a more stable value. In Indonesia, there are futures exchange companies that offer gold investment with an online transaction system (E-Trade). The amount of demand and supply, the rate of inflation, economic conditions, and many more can affect the high and low prices of gold. Due to changes in the conditions above, the price of gold may increase, decrease, or remain constant every day. The price of gold that can go up and down causes the need for gold price predictions so that future gold trading investment prospects can be seen. In this final project, the accuracy of Support Vector Regression will be investigated to find out how accurate it is in predicting gold prices with High, Low, Open, Close, and Volume variables. Based on the calculation of the best RMSE in the study, it was found that the best RMSE was to use a Linear kernel with a C of 35 and using a Y variable dataset of 7.4615. The Support Vector Regression Algorithm can predict quite well, as evidenced by the acquisition of fairly good RMSE results. It is necessary to do a simulation of buying and selling gold based on the prediction results and comparing the advantages of the testing data and the actual data.

Multi-State Load Demand Forecasting Using Hybridized Support Vector Regression Integrated with Optimal Design of Off-Grid Energy Systems—A Metaheuristic Approach

The prediction accuracy of support vector regression (SVR) is highly influenced by a kernel function. However, its performance suffers on large datasets, and this could be attributed to the computational limitations of kernel learning. To tackle this problem, this paper combines SVR with the emerging Harris hawks optimization (HHO) and particle swarm optimization (PSO) algorithms to form two hybrid SVR algorithms, SVR-HHO and SVR-PSO. Both the two proposed algorithms and traditional SVR were applied to load forecasting in four different states of Nigeria. The correlation coefficient (R), coefficient of determination (R2), mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE) were used as indicators to evaluate the prediction accuracy of the algorithms. The results reveal that there is an increase in performance for both SVR-HHO and SVR-PSO over traditional SVR. SVR-HHO has the highest R2 values of 0.9951, 0.8963, 0.9951, and 0.9313, the lowest MSE values of 0.0002, 0.0070, 0.0002, and 0.0080, and the lowest MAPE values of 0.1311, 0.1452, 0.0599, and 0.1817, respectively, for Kano, Abuja, Niger, and Lagos State. The results of SVR-HHO also prove more advantageous over SVR-PSO in all the states concerning load forecasting skills. This paper also designed a hybrid renewable energy system (HRES) that consists of solar photovoltaic (PV) panels, wind turbines, and batteries. As inputs, the system used solar radiation, temperature, wind speed, and the predicted load demands by SVR-HHO in all the states. The system was optimized by using the PSO algorithm to obtain the optimal configuration of the HRES that will satisfy all constraints at the minimum cost.

Noninvasive blood glucose sensing by near-infrared spectroscopy based on PLSR combines SAE deep neural network approach

Near-infrared spectroscopy has been considered as one of the most effective methods for noninvasive blood glucose sensing. Due to the strong scattering of human tissues and the differences among individuals, the relationship between spectral data and blood glucose concentration is nonlinear. Therefore, the linear prediction model has limitations when modeling multiple human samples. The present paper proposes a hybrid model in order to improve the prediction accuracy and versatility of the method, which was based on integrated linear partial least square regression (PLSR) with the nonlinear stacked auto-encoder (SAE) deep neural network. In this work, the diffuse reflectance spectrum of the palm was measured at six different wavelengths in 19 healthy subjects. The prediction results of multiple samples demonstrated that the correlation coefficients of the PLSR-SAE model is improved from 0.3021 to 0.9216 on average, which significantly optimizes the prediction effect compared with the traditional PLSR model. In addition, the prediction accuracy of Support Vector Regression (SVR) model and PLSR-SAE model are 0.8243 and 0.9216 respectively. Furthermore, in Clarke error grid analysis, the PLSR-SAE model could achieve 97.96% of points in A region, which has demonstrated that the prediction accuracy of this noninvasive blood glucose detection method might meets the precision range of clinical laboratory standards. Furthermore, it shows the potential of combining linear and nonlinear regression models for noninvasive prediction of other blood components.

Sparse auto encoder driven support vector regression based deep learning model for predicting network intrusions

The Network Intrusion Detection System (NIDS) assumes a prominent aspect in ensuring network security. It serves better than traditional network security mechanisms, such as firewall systems. The result of the NIDS indicates the enhanced and efficient performance of the algorithms. It is utilized to predict intrusions, and it also has better training times for the algorithms. In this paper, a capable deep learning model using Sparse Auto Encoder (SAE) is proposed. It is a self-taught learning framework. Such a model is a competent unsupervised learning algorithm in reconstructing new feature representation; thus, it diminishes the dimensionality. The SAE requires minimum training time substantially and efficiently enhances the prediction accuracy of Support Vector Regression (SVR) related to attacks. The experiments are administered using the standard intrusion detection dataset NSL-KDD, and therefore, the implementations are performed using python and tensor flow. The proposed model’s effectiveness is estimated with other models viz., the PCA-SVR and SVR models applying prediction metrics such as R2 score, Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), accuracy and also training time. Results validate that the proposed SAE-SVR model has accelerated the training time of SVR and has the edge over the other models weighed in terms of prediction metrics. The model improves the rate of prediction by bringing down the error rates and yields a pioneering research mechanism for predicting the intrusions.

Multiple optimized support vector regression for multi-sensor data fusion of weigh-in-motion system

Weigh-in-motion is an efficient way to manage overload vehicles, and usually utilizes multi-sensor to measure vehicle weight at present. To increase generalization and accuracy of support vector regression (SVR) applied in multi-sensor weigh-in-motion data fusion, three improved algorithms are presented in this paper. The first improved algorithm divides train samples into two sets to construct SVR1 and SVR2, respectively, and then test samples are distributed to SVR1 or SVR2 based on the nearest distance principle. The second improved algorithm calculates the theoretical biases of two training samples closeted to one test sample, and then obtains the bias of the test sample by linear interpolation method. The third improved algorithm utilizes the second improved algorithm to realize adaptive adjustment of biases for SVR1 and SVR2. Five vehicles were selected to conduct multi-sensor weigh-in-motion experiments on the built test platform. According to the obtained experiment data, fusion tests of SVR and three improved algorithms are performed, respectively. The results show that three improved algorithms gradually increase accuracy of SVR with fast operation speed, and the third improved algorithm exhibits the best application prospect in multi-sensor weigh-in-motion data fusion.

Shear force estimation in rough boundaries using SVR method

The accuracy of support vector regression (SVR) procedure in modeling the percentage of shear force carried by walls in a rectangular channel with rough boundaries was investigated. The SVR model is extended, and the more appropriate kernel function and input combination are studied. Finally, the SVR model with an exponential kernel function and three influence parameters was selected as the best SVR model with the lowest error. The output of this more appropriate SVR model is presented as a program. Then, this most appropriate SVR model is compared with three equations presented by other researchers for rough and smooth channels. The SVR model with the highest accuracy and lowest statistical values (RMSE of 0.565) performed the best compared with the other equations.

Feature Selection Using Gram-Schmidt Orthogonalization For Support Vector Regression – A Case Study Of Mortality Rate Prediction Caused By Pneumonia

Feature selection is a technique for finding optimal features among original features by eliminating irrelevant features. Besides improving the learning accuracy and facilitate a better understanding of the model, feature selection may reduce the cost of building, storing and processing models. Recently, a Gram-Schmidt Orthogonalization-based feature selection is proposed for unstructured data. In this paper, we extend this Gram-Schmidt Orthogonalization-based feature selection for structured data. Our simulation shows that this Gram-Schmidt Orthogonalization-based feature selection improves the accuracy of Support Vector Regression in the average of 1.384925% for the case study of the prediction of mortality rates caused by pneumonia.

A method to detect water-injected pork based on bioelectrical impedance technique

The objective of this study was to detect water-injected pork using electrical impedance spectroscopy (EIS) and an artificial intelligence model in a rapid, accurate, and minimally destructive method. Pork loins were injected with water (0%, 3.5%, 7%, 10.5%, 14% and 17.5%, respectively) and physicochemical measurements including cooking loss, pressing loss, color, and textural properties were evaluated after injection. Results indicated that injection of water negatively affected physicochemical characteristics of meat samples (P < 0.05). Water-injected pork had increased cooking loss and pressing loss. There were also significant differences found in color and textural properties. In addition, prediction models for the correlation between impedance magnitude and water injection rate of meat samples were established using support vector regression (SVR) and partial least square regression (PLSR). The results showed that both models performed well at different water injection rate. The model based on SVR obtained optimal performance. From the analysis of meat samples, the accuracy of SVR, which yielded 85.3%, was superior to that of PLSR with accuracy of 84.0%. EIS could be a potential technique for detecting water-injected porcine meat produced in the pork industry.

HazeEst: Machine Learning Based Metropolitan Air Pollution Estimation From Fixed and Mobile Sensors

Metropolitan air pollution is a growing concern in both developing and developed countries. Fixed-station monitors, typically operated by governments, offer accurate but sparse data, and are increasingly being augmented by lower fidelity but denser measurements taken by mobile sensors carried by concerned citizens and researchers. In this paper, we introduce HazeEst—a machine learning model that combines sparse fixed-station data with dense mobile sensor data to estimate the air pollution surface for any given hour on any given day in Sydney. We assess our system using seven regression models and tenfold cross validation. The results show that estimation accuracy of support vector regression (SVR) is similar to decision tree regression and random forest regression, and higher than extreme gradient boosting, multi-layer perceptrons, linear regression, and adaptive boosting regression. The air pollution estimates from our models are validated via field trials, and results show that SVR not only yields high spatial resolution estimates that correspond well with the pollution surface obtained from fixed and mobile sensor monitoring systems, but also indicates boundaries of polluted area better than other regression models. Our results can be visualized using a Web-based application customized for metropolitan Sydney. We believe that the continuous estimates provided by our system can better inform air pollution exposure and its impact on human health.

Regional Forecasting with Support Vector Regressions: The Case of Spain

This study attempts to assess the forecasting accuracy of Support Vector Regression (SVR) with regard to other Artificial Intelligence techniques based on statistical learning. We use two different neural networks and three SVR models that differ by the type of kernel used. We focus on international tourism demand to all seventeen regions of Spain. The SVR with a Gaussian kernel shows the best forecasting performance. The best predictions are obtained for longer forecast horizons, which suggest the suitability of machine learning techniques for medium and long term forecasting.

Life Prediction for Silicon Pressure Sensor Based on SVR

In order to improve the reliability of silicon pressure sensor, life prediction for silicon pressure sensor should be performed. Life prediction for silicon pressure sensor based on support vector regression is proposed in the paper. Grid method is used to determine the parameters of support vector regression in the process of training support vector regression model. Life for silicon pressure sensor under the conditions of different pressures is given in the experimental analysis. The comparison of the errors and mean errors between support vector regression and BP neural network indicates that life prediction accuracy of support vector regression for silicon pressure sensor is higher than that of BP neural network.