RBFNN Method Research Articles

A candidate method for prediction of the non-stationary response of strongly nonlinear systems under wide-band noise excitation

This paper expands the application of the RBFNN method to the prediction of non-stationary responses in strongly nonlinear systems subjected to wide-band noises (WBNs) excitation. Specifically, the SAM is initially employed to deduce the averaged Fokker–Planck–Kolmogorov (FPK) equation. The non-stationary responses of the energy of the system under WBNs are modeled as a one-dimensional Markov diffusion process. This yields a one-dimensional FPK equation that governs the NSRs of the system subject to WBNs. Subsequently, solving the corresponding FPK equation and obtaining the system’s NSRs are accomplished with the RBFNN method. Furthermore, four indicative cases exposed to WBNs are studied to verify the proposed scheme. Comparisons with simulations demonstrate the accuracy and efficiency of the developed technique.

Gaussian Rbfnn Method for Solving Fpk and Bk Equations in Stochastic Dynamical System with Fopid Controller

Gaussian Rbfnn Method for Solving Fpk and Bk Equations in Stochastic Dynamical System with Fopid Controller

Antioxidant Indexes and Immune Function of the Intestinal Flora of Compound Microecological Preparations.

Compound probiotics are biological products made by mixing and cultivating various probiotics according to appropriate composition and ratio. It can not only play the role of a variety of probiotics in composition but also the effect of the ratio is often greater than the effect of the single strain alone. These probiotics are good for the body's immune system in living organisms. They act as immune boosters in living organisms. They are considered the safest form of treatment. In this review, the effects of microecological agents on antioxidant indices and immune function of chicks were studied, and radial basis neural network (RBFNN) was described. Aiming at the deficiency of traditional RBFNN learning strategy, a three-layer RBFNN method based on immune mechanism was proposed. In this paper, the microecological evaluation preparation was added into chicken feed. The study showed that chickens fed with probiotics gained significant body weight at 6 weeks of age. Compared with the control group, the experimental group increased by 15.70%. In each feed/gain ratio, the feed/gain ratio of the experimental group was significantly decreased at 6 weeks of age, which was 30.37% lower than that of the control group. Compared with the control group, the spleen, bursa of Fabricius, and thymus of the experimental group were increased by 63.02%, 38.77%, and 26.82%, respectively, indicating that the compound probiotics could improve the immunity of chickens.

Bitcoin USD Closing Price (BTC-USD) Comparison Using Simple Moving Average And Radial Basis Function Neural Network Methods

Bitcoin is a decentralized electronic money that is not controlled nor insured by a central authority. Because it is still a young system, the price of Bitcoin is extremely unpredictable, making Bitcoin users and investors uneasy. A typical difficulty for investors and traders is predicting the future movement of the value of Bitcoin electronic money based on historical data. Because investors and traders only notice swings in global currency prices and make Bitcoin buy/sell decisions instinctively, they frequently make the erroneous buy/sell decisions. Many investors and traders suffered significant losses as a result of this error. Losses can be reduced by employing an algorithm that predicts the movement of the value of Bitcoin electronic money. Using a comparison of two methodologies, the Simple Moving Average and RBFNN, we will anticipate the closing price of Bitcoin USD (BTC-USD) from January 1, 2021 to January 31, 2021. The results obtained using the simple moving average method MSE = 0.01 percent and MAPE = 36.67 percent, and the results obtained using the RBFNN method MSE = 9.97 x 10-7 and MAPE = 9.97 x 10-5, indicating that the RBFNN method with an accuracy rate of 99.9995 percent is better than the simple moving average method in forecasting the closing price of bitcoin.

Estimating the compressive strength of rectangular fiber reinforced polymer–confined columns using multilayer perceptron, radial basis function, and support vector regression methods

Recently, there has been a tendency to use machine learning (ML)–based methods, such as artificial neural networks (ANNs), for more accurate estimates. This paper investigates the effectiveness of three different machine learning methods including radial basis function neural network (RBNN), multi-layer perceptron (MLP), and support vector regression (SVR), for predicting the ultimate strength of square and rectangular columns confined by various FRP sheets. So far, in the previous study, several experiments have been conducted on concrete columns confined by fiber reinforced polymer (FRP) sheets with the results suggesting that the use of FRP sheets enhances the compressive strength of concrete columns effectively. Also, a wide range of experimental data (including 463 specimens) has been collected in this study for square and rectangular columns, confined by various FRP sheets. The comparison of ML-derived results with the experimental findings, which were in a very good agreement, demonstrated the ability of ML to estimate the compressive strength of concrete confined by FRP; the correlation coefficient (R2) for MLP, RBFNN, and SVR methods was equal to 0.97, 0.97, and 0.90, respectively. Similar accuracy was obtained by MLP and RBFNN, and they provided better estimates for determining the compressive strength of concrete confined by FRP. Also, the results showed that the difference between statistical indicators for training and testing specimens in the RBFNN method was greater than the MLP method, and this difference indicated the poor performance of RBFNN.

An Effective of Ensemble Boosting Learning Method for Breast Cancer Virtual Screening Using Neural Network Model

Radial Based Function Neural Network models (RBFNN) are currently used deep-rooted methods for assessing the stages of diagnosis of chronic diseases. The goals of this research are to suggest a model for the diagnosis of breast cancer, and to be able to estimate the stages of development of pre-malignant breast tumors. The significance of the study is to develop an integrated RBF neural network with ensemble features using the boosting method. The importance of the ensemble boosting method is to generate a sequence of models to achieve more precise predictions. One of the ensemble boosting advantages is that it can take longer to build and to score than a RBF NN standard model. By using ensemble boosting, the accuracy of breast tumor diagnosis increased and thus it became easier to know the stage of the tumor, and whether it was malignant or benign. This will help doctors to select appropriate treatment for each tumor stage, consequently leading to the salvage of cancer patients with this type of tumor. The suggested RBFNN method was examined on the different type of UCI breast cancer datasets. The general diagnosis accuracy based on 10-fold cross validation using RBFNN method obtained 97.4%, 98.4%, 97.7% and 97.0% for the WBC, BCD, BCP, and WBCD UCI datasets respectively. The effectiveness of the proposed method was confirmed by comparing accuracy improvement both before and after using ensemble boosting, and it was found to be more accurate compared with other breast cancer diagnosis methods such as Logistic Regression (91.5%), KNN (96%), SVM (89%), Decision tree (95.13), CNN (97.66%), and Naive Bayes (91%).

Improved Imaging Resolution of Electrical Impedance Tomography Using Artificial Neural Networks for Image Reconstruction.

Electrical impedance tomography (EIT) is a noninvasive and non-radiative medical imaging technique based on detecting the inhomogeneous electrical properties of the tissue. The inverse problem of EIT is a highly nonlinear ill-posed problem, which is the main reason that affects image quality. Our goal is to solve the EIT inverse problem using the nonlinear mapping properties of artificial neural networks (ANNs) and convolutional neural networks (CNNs). In this paper, the adaptive moment estimation (ADAM) optimization method and mean-square-error (MSE) function are used to train an ANN to solve the inverse problem and a CNN to process the ANN image. The networks are trained on datasets of simulated data, and tested on datasets of simulated data and experimental data. Results for time-difference EIT (td-EIT) images are presented using simulated EIT data from EIDORS and experimental EIT data from our EIT systems. The results are used to compare the proposed method with the one-step Gauss-Newton linear method and RBFNN method. The proposed method offers improved resolution (RES), low position error (PE) and excellent artefact removal compared to the existing methods. The experimental results show that our method can improve the RES by 50 to 70 percent and reduce the PE by 60 to 70 percent. The improvements in RES and processing speed are essential for clinical EIT measurement of dynamic physiological processes.

Residual range estimation for battery electric vehicle based on radial basis function neural network

The accurate residual range estimation of a battery electric vehicle (BEV) can alleviate the driver’s range anxiety and improve driving safety. In this study, the residual range estimation is considered a nonlinear system involving battery factors and a variety of vehicle working status factors. Given that the radial basis function neural network (RBF NN) performs well in approximating nonlinear systems, this study proposes a RBF NN method to estimate the residual range of BEV. The contribution analysis method is used to simplify the input layer of RBF NN and enhance the real-time performance of estimation. Then, the residual range is estimated by RBF NN using historical data newly collected at a fixed time in current discharge process. Some experimental data are from operational BEVs in Beijing, China. Experimental results of different types demonstrate that all errors are comparatively small and within the engineering limit.

Stock price prediction based on chaotic hybrid particle swarm optimisation-RBF neural network

The stock market is an important part of the capital market, which plays a significant role in optimising capital allocation, financing and increasing the value of assets and other areas. Hence, the correct model for estimating and predicting the stock price has a very important practical significance to provide investors with investment decision reference. In this paper, a novel chaotic hybrid PSO-based RBF neural network model (CHPSO-RBFNN) has been proposed for forecasting the stock price, which can effectively prevent the RBF neural network from the local minimum trap and provide great learning ability. The presented methodology was tested with stock 601998, and the results showed that CHPSO-RBFNN can improve the prediction of accuracy and a high efficient and accurate stock prediction model compared to the traditional RBFNN and PSO-RBFNN methods.

Stock price prediction based on chaotic hybrid particle swarm optimization-RBF neural network

The stock market is an important part of the capital market, which plays a significant role in optimising capital allocation, financing and increasing the value of assets and other areas. Hence, the correct model for estimating and predicting the stock price has a very important practical significance to provide investors with investment decision reference. In this paper, a novel chaotic hybrid PSO-based RBF neural network model (CHPSO-RBFNN) has been proposed for forecasting the stock price, which can effectively prevent the RBF neural network from the local minimum trap and provide great learning ability. The presented methodology was tested with stock 601998, and the results showed that CHPSO-RBFNN can improve the prediction of accuracy and a high efficient and accurate stock prediction model compared to the traditional RBFNN and PSO-RBFNN methods.

Utilizing RBF-NN and ANFIS Methods for Multi-Lead ahead Prediction Model of Evaporation from Reservoir

Evaporation as a major meteorological component of the hydrologic cycle plays a key role in water resources studies and climate change. The estimation of evaporation is a complex and unsteady process, so it is difficult to derive an accurate physical-based formula to represent all parameters that effect on estimate evaporation. Artificial intelligence-based methods may provide reliable prediction models for several applications in engineering. In this research have been introduced twelve networks with the RBF-NN and ANFIS methods. These models have applied to prediction daily evaporation at Layang reservoir, located in the southeast part of Malaysia. The used meteorological data set to develop the models for prediction daily evaporation rate from water surface for Layang reservoir includes daily air temperature, solar radiation, pan evaporation, and relative humidity that measured at a case study for fourteen years. The obtained result denote to the superiority of the RBF-NN models on the ANFIS models. A comparison of the model performance between RBF-NN and ANFIS models indicated that RBF-NN method presents the best estimates of daily evaporation rate with the minimum MSE 0.0471 , MAE 0.0032, RE and maximum R2 0.963.

Nonlinear inversion of electrical resistivity imaging using pruning Bayesian neural networks

Conventional artificial neural networks used to solve electrical resistivity imaging (ERI) inversion problem suffer from overfitting and local minima. To solve these problems, we propose to use a pruning Bayesian neural network (PBNN) nonlinear inversion method and a sample design method based on the K-medoids clustering algorithm. In the sample design method, the training samples of the neural network are designed according to the prior information provided by the K-medoids clustering results; thus, the training process of the neural network is well guided. The proposed PBNN, based on Bayesian regularization, is used to select the hidden layer structure by assessing the effect of each hidden neuron to the inversion results. Then, the hyperparameter αk, which is based on the generalized mean, is chosen to guide the pruning process according to the prior distribution of the training samples under the small-sample condition. The proposed algorithm is more efficient than other common adaptive regularization methods in geophysics. The inversion of synthetic data and field data suggests that the proposed method suppresses the noise in the neural network training stage and enhances the generalization. The inversion results with the proposed method are better than those of the BPNN, RBFNN, and RRBFNN inversion methods as well as the conventional least squares inversion.

Real-time Pose Measurement of Parallel Robot Based on GRNN

The real-time pose measurement of parallel robot helps to achieve the closed loop pose control and improve the control and operating performance of parallel robot. But it is difficult to implement the real-time pose measurement directly. In order to solve the pose measurement problem of a 6-DOF parallel robot, the kinematics analysis of the parallel robot is made, and a Generalized Regression Neural Network which has fast convergence and strong nonlinear mapping ability is established by setting the desired pose and its inverse kinematics results as the neural network training samples to implement the map of parallel robot from the joint variable space to the work variable space. Finally, the real-time pose measurement of parallel robot is achieved by using the trained neural network and the actual motion states of the active joints easily detected. The simulation experiment results show that the method of measuring the parallel robot pose based on the GRNN has the faster convergence rate and higher measurement accuracy than those of the BPNN and RBFNN methods. The research establishes the basis for the direct closed control of parallel robot pose. DOI: http://dx.doi.org/10.11591/telkomnika.v11i5.2455

Application of DG-RBFNN in Power Flow Calculation of Distribution Network with DG

With the injection of distributed generation (DG) into distribution network, complexity and uncertainty of power flow in distribution network follow. In order to calculate power flow more simply and accurately, considering the characteristics of distribution network with DG and the flexible applicability of RBF neural network, a special adaptive dynamic clustering RBFNN (DG-RBFNN) method, which clusters the input samples only according to the parameters associated with DG, has been used in this paper. Therefore, the results are more approximate to the real condition and the calculation process is simpler compared with conventional back/forward (B/F) method meanwhile the calculation scale is also relatively smaller compared with ordinary adaptive dynamic clustering RBFNN. Finally, according to a 21-bus 66 kV distribution network of Shenyang, Liaoning province simulation experiment, the availability of DG-RBFNN method is proved.

Diffusion coefficient prediction of acids in water at infinite dilution by QSPR method

In this article, quantitative structure–property relationship (QSPR) models were developed for estimation of diffusion coefficients (DCs) of acids at infinite dilution in water. These models were obtained based on molecular descriptors of a set of 65 compounds at 298.15 K. Five date splits are randomly extracted from this dataset. Using a genetic algorithm (GA) variable-selection approach, five molecular descriptors in each split were selected from a set of thousand of them. To analyze the nonlinear behavior of these molecular descriptors, Adaptive neuro-fuzzy inference system (ANFIS) and radial basis function neural network (RBF NN) were used. These models were trained on 52 arbitrary sets of the data, and the remaining 13 sets were employed to evaluate the proposed models. Genetic function approximation (GFA) method could predict DCs of test dataset with high value of squared correlation coefficients of 0.984. QSPR model even gave superior predictions. ANFIS and RBF NN methods when implemented in QSPR model provided the squared correlation coefficients of 0.989 and 0.987, respectively. It is therefore suitable to be used in the proposed QSPR models, especially when applied with ANFIS method, to obtain DC of acids at infinite dilution in water instead of other expensive and complicated experiments.

RBF와 LVQ 인공신경망을 이용한 요(尿) 딥스틱 선별검사에서의 요로감염 분류

Dipstick urinalysis is used as a routine test for a screening test of UTI (urinary tract infection) in primary practice because urine dipstick test is simple. The result of dipstick urinalysis brings medical professionals to make a microscopic examination and urine culture for exact UTI diagnosis, therefore it is emphasized on a role of screening test. The objective of this study was to the classification between UTI patients and normal subjects using hybrid neural network classifier with enhanced clustering performance in urine dipstick screening test. In order to propose a classifier, we made a hybrid neural network which combines with RBF layer, summation & normalization layer and L VQ artificial neural network layer. For the demonstration of proposed hybrid neural network, we compared proposed classifier with various artificial neural networks such as back-propagation, RBFNN and PNN method. As a result, classification performance of proposed classifier was able to classify 95.81% of the normal subjects and 83.87% of the UTI patients, total average 90.72% according to validation dataset. The proposed classifier confirms better performance than other classifiers. Therefore the application of such a proposed classifier expect to utilize telemedicine to classify between UTI patients and normal subjects in the future.

The transfer function model for dynamic response of wet cooling coils

Abstract This paper mainly concerned about the dynamic response model of wet cooling coils that is developed by the Laplace transform method. The theoretic equations are firstly established based on the theory of energy conservation. Then, the transfer functions on the transient responses of wet cooling coils have been deduced using the method of Laplace transform. The transfer functions reveal the dynamic relationships between the inlet variables and the outlet ones of the cooling coils. Partial-fraction method and Newton–Raphson method are both used in the inversion of the transfer functions from the s -domain to τ -domain. To make the dynamic model of wet cooling coils more adaptive, RBFNN method is employed to determine the coefficients of heat and mass transfer. Experiments have been done and manifested that the coefficients of heat and mass transfer by RBFNN will be of great value to the validity of the transient response model of wet cooling coils in this study.

DIRECTION OF ARRIVAL AND STATE OF POLARIZATION ESTIMATION USING RADIAL BASIS FUNCTION NEURAL NETWORK (RBFNN)

A neural network architecture is applied to the problem of direction of arrival (DOA) and state of polarization estimation using a uniform circular cross and tri-crossed-dipoles antenna array. A three layer radial basis function network (RBFN) is trained with input output pairs. The network is then capable of estimating DOA not included in the training set through generalization and the corresponding state of polarization. This approach reduces the extensive computations required by conventional super resolution algorithms such as MUSIC and is easier to implement in real-time applications. The results suggest that the performance of the RBFNN method approaches the exact values. In real time, fast convergence rates of neural networks will allow the array to track mobile sources.

RBFNN Method of Calculating Energy Losses of Power Distribution Systems

RBFNN Method of Calculating Energy Losses of Power Distribution Systems

Application of support vector machine (SVM) for prediction toxic activity of different data sets

As a new method, support vector machine (SVM) were applied for prediction of toxicity of different data sets compared with other two common methods, multiple linear regression (MLR) and RBFNN. Quantitative structure–activity relationships (QSAR) models based on calculated molecular descriptors have been clearly established. Among them, SVM model gave the highest q 2 and correlation coefficient R. It indicates that the SVM performed better generalization ability than the MLR and RBFNN methods, especially in the test set and the whole data set. This eventually leads to better generalization than neural networks, which implement the empirical risk minimization principle and may not converge to global solutions. We would expect SVM method as a powerful tool for the prediction of molecular properties.