Radial Basis Function Neural Network Method Research Articles

Image Reconstruction for Electrical Impedance Tomography Using Radial Basis Function Neural Network Based on Hybrid Particle Swarm Optimization Algorithm

A Hybrid Particle Swarm Optimization (HPSO) algorithm is proposed to optimize the Radial Basis Function Neural Network (RBFNN) for the image reconstruction of Electrical Impedance Tomography (EIT). This algorithm combines the non-linear approximation ability of RBFNN with the mutation ability of HPSO based on the simulated annealing, which has advantages of the strong global search ability and high robustness. A non-linear mapping relationship model between the boundary voltage and the conductivity distribution is established through HPSO-RBFNN to realize the image reconstruction, skipping the solution of the Jacobian matrix. The Structural similarity (SS) and correlation coefficient (CC) are used as the evaluation criteria for the quality of the reconstructed images in the simulations and experiments. Firstly, the HPSO-RBFNN method is compared with the Landweber and RBFNN methods through a single-target simulation, which verifies the accuracy of the image reconstruction. Secondly, the Gaussian noise is added to the ideal simulation data, the results show an anti-noise capability of the proposed method. Then, the multi-target simulation is conducted using the single-target training model, which verifies the generalization ability of the proposed method. Finally, the experiments with multiple fish eggs are performed to further verify the effectiveness of the HPSO-RBFNN method. The results show that the method proposed in this article has higher SS and CC, which indicates that the HPSO-RBFNN method has a higher accuracy.

A High Precision Adaptive Back-Stepping Control Method for Morphing Aircraft Based on RBFNN Method

To overcome the uncertainties of the nonlinear model of a morphing aircraft, this paper presents a high-precision adaptive back-stepping control method based on the radial basis function neural network (RBFNN). Firstly, based on the analysis of static and dynamic aerodynamic parameters of the morphing aircraft, its nonlinear control law is designed by using the conventional back-stepping method. The RBFNN is introduced to approximate online the uncertain terms of the nonlinear control law so as to improve its robustness. The robust term is designed to eliminate the approximation error caused by the RBFNN. Secondly, the tracking differentiator is designed through solving the virtual control variables, thus solving the "differential expansion" problem existing in the traditional back-stepping method. The Lyapunov stability analysis proves that our method can ensure that the tracking error of a closed-loop system converges finally and that its signals are uniformly bounded. Finally, the digital simulation model of the morphing aircraft is established with the MATLAB/Simulink; our method is compared with the conventional back-stepping control method. The simulation results show that our method has a higher control precision and stronger robustness.

PERBAIKAN KONTRAS CITRA MAMMOGRAM PADA KLASIFIKASI KANKER PAYUDARA BERDASARKAN FITUR GRAY-LEVEL CO-OCCURRENCE MATRIX

In this era to recognize breast tumors can be based on mammogram images. This method will expedite the process of recognition and classification of breast cancer. This research was conducted classification techniques of breast cancer using mammogram images. The proposed model targets classification studies for cases of malignant, and benign cancer. The research consisted of five main stages, preprocessing, histogram equalization, convolution, feature extraction, and classification. For preprocessing cropping the image using region of interest (ROI), for convolution, median filter and histogram equalization are used to improve image quality. Feature extraction using Gray-Level Co-Occurrence Matrix (GLCM) with 5 features, entropy, correlation, contrast, homogeneity, and variance. The final step is the classification using Radial Basis Function Neural Network (RBFNN) and Support Vector Machine (SVM). Based on the hypotheses that have been tested and discussed, the accuracy for RBFNN is 86.27%, while the accuracy for SVM is 84.31%. This shows that the RBFNN method is better than SVM in distinguishing types of breast cancer. These results prove the process of improving image construction using histogram equalization and the median filter is useful in the classification process.

Experimental Test and Prediction Model of Soil Thermal Conductivity in Permafrost Regions

Soil thermal conductivity is a dominant parameter of an unsteady heat-transfer process, which further influences the stability and sustainability of engineering applications in permafrost regions. In this work, a laboratory test for massive specimens is performed to reveal the distribution characteristics and the parameter-influencing mechanisms of soil thermal conductivity along the Qinghai–Tibet Engineering Corridor (QTEC). Based on the measurement data of 638 unfrozen and 860 frozen soil specimens, binary fitting, radial basis function (RBF) neural network and ternary fitting (for frozen soils) prediction models of soil thermal conductivity have been developed and compared. The results demonstrate that, (1) particle size and intrinsic heat-conducting capacity of the soil skeleton have a significant influence on the soil thermal conductivity, and the typical specimens in the QTEC can be classified as three clusters according to their thermal conductivity probability distribution and water-holding capacity; (2) dry density as well as water content sometimes does not have a strong positive correlation with thermal conductivity of natural soil samples, especially for multiple soil types and complex compositions; (3) both the RBF neural network method and ternary fitting method have favorable prediction accuracy and a wide application range. The maximum determination coefficient (R2) and quantitative proportion of relative error within ±10% ( P ± 10 % ) of each prediction model reaches up to 0.82, 0.88, 81.4% and 74.5%, respectively. Furthermore, because the ternary fitting method can only be used for frozen soils, the RBF neural network method is considered the optimal approach among all three prediction methods. This study can contribute to the construction and maintenance of engineering applications in permafrost regions.

Investigating the Predictive Performance of Gaussian Process Regression in Evaluating Reservoir Porosity and Permeability

In this paper, a new predictive model based on Gaussian process regression (GPR) that does not require iterative tuning of user-defined model parameters has been proposed to determine reservoir porosity and permeability. For this purpose, the capability of GPR was appraised statistically for predicting porosity and permeability of the southern basin of the South Yellow Sea using petrophysical well log data. Generally, the performance of GPR is deeply reliant on the type covariance function utilized. Therefore, to obtain the optimal GPR model, five different kernel functions were tested. The resulting optimal GPR model consisted of the exponential covariance function, which produced the highest correlation coefficient (R) of 0.85 and the least root mean square error (RMSE) of 0.037 and 6.47 for porosity and permeability, respectively. Comparison was further made with benchmark methods involving a back propagation neural network (BPNN), generalized regression neural network (GRNN), and radial basis function neural network (RBFNN). The statistical findings revealed that the proposed GPR is a powerful technique and can be used as a supplement to the widely used artificial neural network methods. In terms of computational speed, the GPR technique was computationally faster than the BPNN, GRNN, and RBFNN methods in estimating reservoir porosity and permeability.

Performance of machine learning methods in diagnosing Parkinson's disease based on dysphonia measures.

Parkinson's disease (PD) is a widespread degenerative syndrome that affects the nervous system. Its early appearing symptoms include tremor, rigidity, and vocal impairment (dysphonia). Consequently, speech indicators are important in the identification of PD based on dysphonic signs. In this regard, computer-aided-diagnosis systems based on machine learning can be useful in assisting clinicians in identifying PD patients. In this work, we evaluate the performance of machine learning based techniques for PD diagnosis based on dysphonia symptoms. Several machine learning techniques were considered and trained with a set of twenty-two voice disorder measurements to classify healthy and PD patients. These machine learning methods included linear discriminant analysis (LDA), k nearest-neighbors (k-NN), naïve Bayes (NB), regression trees (RT), radial basis function neural networks (RBFNN), support vector machine (SVM), and Mahalanobis distance classifier. We evaluated the performance of these methods by means of a tenfold cross validation protocol. Experimental results show that the SVM classifier achieved higher average performance than all other classifiers in terms of overall accuracy, G-mean, and area under the curve of the receiver operating characteristic plot. The SVM classifier achieved higher performance measures than the majority of the other classifiers also in terms of sensitivity, specificity, and F-measure statistics. The LDA, k-NN and RT achieved the highest average precision. The RBFNN method yielded the highest F-measure.; however, it performed poorly in terms of other performance metrics. Finally, t tests were performed to evaluate statistical significance of the results, confirming that the SVM outperformed most of the other classifiers on the majority of performance measures. SVM is a promising method for identifying PD patients based on classification of dysphonia measurements.

Improving prediction accuracy of cooling load using EMD, PSR and RBFNN

To increase the accuracy for the prediction of cooling load demand, this work presents an EMD (empirical mode decomposition)-PSR (phase space reconstruction) based RBFNN (radial basis function neural networks) method. Firstly, analyzed the chaotic nature of the real cooling load demand, transformed the non-stationary cooling load historical data into several stationary intrinsic mode functions (IMFs) by using EMD. Secondly, compared the RBFNN prediction accuracies of each IMFs and proposed an IMF combining scheme that is combine the lower-frequency components (called IMF4-IMF6 combined) while keep the higher frequency component (IMF1, IMF2, IMF3) and the residual unchanged. Thirdly, reconstruct phase space for each combined components separately, process the highest frequency component (IMF1) by differential method and predict with RBFNN in the reconstructed phase spaces. Real cooling load data of a centralized ice storage cooling systems in Guangzhou are used for simulation. The results show that the proposed hybrid method outperforms the traditional methods.

Solving medical classification problems with RBF neural network and filter methods

This paper evaluates classification accuracy of radial basis function (RBF) neural network and filter methods for feature selection in medical datasets. To improve the diagnostic procedure in the daily routine and to avoid wrong diagnosis, machine learning methods can be used. Diagnosis of tumours, heart disease, hepatitis, liver and Parkinson's diseases are a few of the medical problems which we have used in artificial neural networks. The main objective of this paper is to show that it is possible to improve the performance of the system for inductive learning rules with RBF neural network for medical classification problems, using the filter methods for feature selections. The aim of this research is also to present and compare different algorithm approach for the construction system that learns from experience and makes decisions and predictions and reduce the expected number or percentage of errors.

Enhanced RBF neural network model for time series prediction of solar cells panel depending on climate conditions (temperature and irradiance)

A radial basis function neural network is an effective technique for function approximation and prediction. It has been used in many applications in the real world; one of them is the time series prediction which is a relatively complex problem. In this paper, we propose an enhanced radial basis function neural network (RBFNN) model that depends on the standard RBF built-in MATLAB (newrb). The enhancement on newrb depends on the use of intelligent algorithms like K-means clustering, K-nearest neighbor, and singular value decomposition, to optimize the centers c, radii r, and weights w of the RBFNN. These algorithms replace the mathematical calculation used to find these parameters in newrb. The proposed enhanced model is applied to predict the solar cells energy production in Palestine using already installed solar panels in Jericho city. Solar irradiance and daily temperature are used as an input training data set for the proposed model, with the real output power of (2015) as the training supervisor. The model is applied to predict the output power within 1 month and for 1 year. Finally, a power output equation was optimized to calculate the solar energy depending on the irradiance and temperature with an acceptable accuracy. The experimental results show that the enhanced model performs more precisely than the traditional RBFNN and multilayer perceptron neural network methods, with low mean square error of relatively few neurons on the hidden layer.

The Optimal Confidence Intervals for Agricultural Products’ Price Forecasts Based on Hierarchical Historical Errors

With the levels of confidence and system complexity, interval forecasts and entropy analysis can deliver more information than point forecasts. In this paper, we take receivers’ demands as our starting point, use the trade-off model between accuracy and informativeness as the criterion to construct the optimal confidence interval, derive the theoretical formula of the optimal confidence interval and propose a practical and efficient algorithm based on entropy theory and complexity theory. In order to improve the estimation precision of the error distribution, the point prediction errors are STRATIFIED according to prices and the complexity of the system; the corresponding prediction error samples are obtained by the prices stratification; and the error distributions are estimated by the kernel function method and the stability of the system. In a stable and orderly environment for price forecasting, we obtain point prediction error samples by the weighted local region and RBF (Radial basis function) neural network methods, forecast the intervals of the soybean meal and non-GMO (Genetically Modified Organism) soybean continuous futures closing prices and implement unconditional coverage, independence and conditional coverage tests for the simulation results. The empirical results are compared from various interval evaluation indicators, different levels of noise, several target confidence levels and different point prediction methods. The analysis shows that the optimal interval construction method is better than the equal probability method and the shortest interval method and has good anti-noise ability with the reduction of system entropy; the hierarchical estimation error method can obtain higher accuracy and better interval estimation than the non-hierarchical method in a stable system.

Comparison of genetic programming and radial basis function neural network for open-channel junction velocity field prediction

The ability to accurately predict the velocity field in open-channel junctions has a significant impact on the designing process of many hydraulic structures, such as irrigation and drainage channels and sewer networks. The gene expression programming (GEP) and radial basis function neural network (RBF-NN) methods are developed in order to find a continuous spatial velocity description using discrete experimental measurements. By using the coordinates of each point and the tributary-to-main-discharge ratio of an open-channel junction, seven different input combinations are investigated. To find the optimum GEP model, various mathematical and linking functions are studied. The RBF-NN models are developed with various numbers of hidden layer nodes and amounts of spread. A comparison of the results indicates that both GEP and RBF-NN methods can accurately simulate flow in junctions. However, the GEP with root-mean-squared error (RMSE) of 0.2361 performs better than RBF-NN with RMSE of 0.2590. Owing to the higher accuracy and explicit output equation of the GEP, this method can be used in practical situations of predicting open-channel characteristics.

A Neural-Network-Based Beamformer for Phased Array Weather Radar

Phased array weather radars, particularly with high temporal resolution, essentially need a robust and fast beamformer to accurately estimate precipitation profiles such as reflectivity and Doppler velocity. In this paper, we introduce a neural-network-based beamformer to address this problem. In particular, the optimum weight vector is computed by modeling the problem as a three-layer radial basis function neural network (RBFNN), which is trained with I/O pairs obtained from the optimum Wiener solution. The RBFNN was chosen because of its characteristic of accurate approximation and good generalization, and its robustness against interference and noise. The proposed RBFNN beamforming method is compared with traditional beamforming methods, namely, Fourier beamforming (FR), Capon beamforming, and the flower pollination algorithm (FPA), which is a recently proposed nature-inspired optimization algorithm. It is shown that the RBFNN approach has nearly optimal performance in various precipitation radar signal simulations relative to the rival methods. The validity of the RBFNN beamformer is demonstrated by using real weather data collected by the phased array radar (PAR) at Osaka University, and compared with, in addition to the FR and FPA methods, the minimum mean square error beamforming method. It is shown that the RBFNN method estimates the reflectivity of the PAR at Osaka University with less clutter level than those of the other three methods.

Study on a Novel Hybrid Intelligent Fault Diagnosis Method Based on Improved DE and RBFNN

The radial basis function neural network (RBFNN) is a great potential artificial intelligence technology and can effectively realize the fault diagnosis for small sample and nonlinear problem. But the parameters of RBFNN model seriously affects the generalization ability and diagnosis accuracy on the great extent. So an improved differential evolution algorithm based on dynamic adaptive adjustment strategy is proposed to optimize the parameters of RBFNN model for obtaining the optimal RBFNN (DASDERBFNN) method. Then the proposed DASDERBFNN method is used to construct a new fault diagnosis (DSDRBFNFD) method. In the DSDRBFNFD method, the dynamic adaptive adjustment strategy is used to adaptively adjust the crossover probability (CR ) value according to the fitness value of current individual in the population for obtaining the improved DE(DASDE) algorithm. Then the selection of parameters in the RBFNN is regarded as a combination optimization of parameters in order to establish the objective function of combination optimization. The DASDE algorithm is used to search for the optimal value of objective function to obtain the better parameter optimization of the RBFNN (DASDERBFNN), which is applied in the fault diagnosis for constructing a new fault diagnosis (DSDRBFNFD) method. Finally, the proposed DSDRBFNFD method is used to diagnose the fault of the cylinder of the engine in order to validate the diagnosis effectiveness of the DSDRBFNFD method. The experiment results show that the proposed DSDRBFNFD method can obtain the higher accuracy of fault diagnosis and is effective fault diagnosis for the engine.

Multimodel Approach Using Neural Networks and Symbolic Regression to Combine the Estimated Discharges of Rainfall-Runoff Models

The aim of this study is to compare the performance of a symbolic regression combination method based on gene expression programming (GEP) with different neural network combination methods when used in the development of multimodel systems. The two different neural network combination methods used in this study are the multilayer perceptron neural network (MLPNN) and the radial basis function neural network (RBFNN). The methods were used to combine the results from different types of rainfall-runoff models to test the multimodel combination system in catchments located in Thailand and New Zealand. Comparison of the results revealed that the GEP performed better than neural network methods in the case of the catchment located in New Zealand. Nevertheless, the RBFNN method outperformed the GEP and the MLPNN combination method in the case of the catchment located in Thailand. However, which combination method produces better results in the multimodel combination is not conclusive. The results suggest that the selection of the best combination method to be used in conjunction with the multimodel approach may depend on the catchment type.

Intelligent Determination of a Battery Energy Storage System Size and Location Based on RBF Neural Networks for Microgrids

As an installation of Battery Energy Storage Systems (BESS) at a non-optimum size and location in a microgrid can cause stability issues, increase in cost, system losses and larger BESS size. Thus, the optimum size and location of BESS is a main challenge for combining BESS into a microgrid. For this reason, this paper proposes a novel method to evaluate an optimal size and location of the BESS using the Radial Basis Function Neural Networks (RBFNN). The proposed method consists of a two-stage based on an optimum size process and optimum location process. In the first stage, the optimal size of the BESS is evaluated using the RBFNN based on frequency and voltage control so that frequency and voltage of the microgrid can return to nominal values under the sudden changes in the microgrid. In the second stage, the optimal location of the BESS is determined using the RBFNN based on minimizing power losses in the microgrid. With a suitable RBFNN training, the results indicate that the proposed RBFNN method can achieve the superb performances in predicting the optimum size and location of the BESS with a minor change compared to the measured data based on the simulation. Therefore, it is obvious that if BESS is located at the optimum location and has the optimum size, it can save an enormous amount of power in a microgrid and can avoid the microgrid from instability and system collapse.

Permeate Flux Measurement and Prediction of Submerged Membrane Bioreactor Filtration Process Using Intelligent Techniques

Recently, membrane technology has become more attractive particularly in solid-liquid separation process. Membrane bioreactor (MBR) has found to be a reliable technology to replace the conventional activated sludge (CAS) process for water and wastewater treatment by adopting membrane filtration technology and bioreactor. However, numerous drawbacks arise when using membrane which includes high maintenance cost and fouling problem. An optimal MBR plant operation is needed to be determined in order to reduce fouling and at the same time reduce the cost of running the MBR. It is crucial to have a reliable MBR filtration prediction that can measure and predict the filtration dynamic performance especially the effect of fouling to the filtration and cleaning operations. With this prediction tool, suitable action can be taken to improve the operation in order to find the optimum setting of the filtration process. This paper presents the permeate flux measurement and prediction development for submerged membrane filtration process. Three input filtration parameters were used to predict the permeate flux in the filtration process. This work employed feed forward artificial neural network (FFNN) and radial basis function neural network (RBFNN) for the prediction purpose. The permeate flux prediction method was developed using operation settings such as aeration airflow, suction pump voltage and transmembrane pressure (TMP) under schedule relaxation condition. The result shows that FFNN method gives better performance compared with RBFNN method in terms of accuracy and reliability.

Wearing prediction of stellite alloys based on opposite degree algorithm

In order to predict the wearing of stellite alloys, the related methods of rare metals data processing were discussed. The method of opposite degree (OD) algorithm was put forward to predict the wearing of stellite alloys. OD algorithm is based on prior numerical data, posterior numerical data and the opposite degree between numerical forecast data. To compare the performance of predicted results based on different algorithms, the back propagation (BP) and radial basis function (RBF) neural network methods were introduced. Predicted results show that the relative error of OD algorithm is smaller than those of BP and RBF neural network methods. OD algorithm is an effective method to predict the wearing of stellite alloys and it can be applied in practice.

Spatial Distribution Characteristics and Hourly Forecast of PM2.5 Pollution in Summer in Beijing

The fine particulate matter (PM2.5) has become the primary pollutant of atmospheric environment of Beijing now, therefore to know the spatial distribution characteristics and to make accurate and refined forecast of PM2.5 is very important. In this study, the spatial distribution characteristics of PM2.5 concentration in summer in Beijing was derived by using Kriging interpolation method based on the data from 30 air quality monitoring stations all over Beijing during June, July and August 2014. In general, the PM2.5 pollution is higher in the south and east, and is lower in the north and west, but when there is south or southeast wind blowing, the situation could be totally the opposite. Based on the hourly PM2.5 concentration data from one representative air quality monitoring station and the meteorological data from a nearby meteorological station during June, July and August 2014, the hourly PM2.5 concentration was forecasted up to 168 hours ahead by using BP (Error-back propagation) and RBF (Radial basis function) neural network methods. The results show RBF neural network method is more efficient, the curve trends of the forecasted values are similar with the curve trends of the monitored values and the forecasted values have significant linear relationship with the monitored values, which demonstrates the possibility of hourly forecast for PM2.5 pollution.

Deception detecting from speech signal using relevance vector machine and non-linear dynamics features

The novel method of deception detecting based on speech signal is proposed in this study. Extracting prosodic and non-linear dynamics (NLD) feature sets from speech signal and applying relevance vector machine (RVM) classification method is the primary target of this paper. Here, the sustained speaker-depended phonation samples of deception and non-deception were applied. In this paper, 30 prosodic features and 18 NLD features were selected which show significant correlations to deception state. Moreover, the RVM classification model based on sparse bayesian learning (SBL) was introduced which is a bayesian extension of the support vector machine (SVM) and not be restricted by the Mercer׳s condition. In the experiments, the deception corpus of Soochow University is exploited to test our approach. Firstly, the experiment of feature performance test was carried out. It is demonstrated that the combination of prosodic and NLD features is the most effectively method for detecting deceptive speech. Secondly, the property of RVM classification model is measured. The experiment results show that RVM technology requires much fewer basic functions and demands much less decision time than the SVM algorithm. Furthermore, during the test of correct classification accuracy of RVM, it shows that RVM is much higher than the classical model of SVM and radial basis function neural network (RBFNN). Finally, the robustness test of RVM is processed and received better performance than SVM and RBFNN methods when the signal to noise ratio (SNR) falls. In general, the proposed deception detecting method based on the combined features and RVM classifier is novel, convenient and practical, moreover, it achieves higher classification accuracy, less detection time, property of generalization, and the strong robustness.

Laser gyro temperature compensation using modified RBFNN.

To overcome the effect of temperature on laser gyro zero bias and to stabilize the laser gyro output, this study proposes a modified radial basis function neural network (RBFNN) based on a Kohonen network and an orthogonal least squares (OLS) algorithm. The modified method, which combines the pattern classification capability of the Kohonen network and the optimal choice capacity of OLS, avoids the random selection of RBFNN centers and improves the compensation accuracy of the RBFNN. It can quickly and accurately identify the effect of temperature on laser gyro zero bias. A number of comparable identification and compensation tests on a variety of temperature-changing situations are completed using the multiple linear regression (MLR), RBFNN and modified RBFNN methods. The test results based on several sets of gyro output in constant and changing temperature conditions demonstrate that the proposed method is able to overcome the effect of randomly selected RBFNN centers. The running time of the method is about 60 s shorter than that of traditional RBFNN under the same test conditions, which suggests that the calculations are reduced. Meanwhile, the compensated gyro output accuracy using the modified method is about 7.0 × 10−4 °/h; comparatively, the traditional RBFNN is about 9.0 × 10−4 °/h and the MLR is about 1.4 × 10−3 °/h.