Learning Radial Basis Function Network Research Articles

Learning Radial Basis Function Networks with the Trust Region Method for Boundary Problems

Learning Radial Basis Function Networks with the Trust Region Method for Boundary Problems

Application of artificial neural networks for viscosity of crude oil-based nanofluids containing oxides nanoparticles

In this research, the effect of general performances of radial basis function (RBF) method of artificial neural networks (ANN) with laboratory data on NiO, WO3, TiO2, ZnO and FeO3 nanoparticles in different temperatures and mass fractions on viscosity of crude oil were studied. The morphology and nanoparticles stability were analyzed with the DLS and TEM analysis. The results showed that the average nanoparticles diameter ranged from 10 to 40 nm for different oxide nanoparticles. For learning RBF networks, the major method for calculating isotropic Gaussian basis functions span for RBF networks containing special algorithm were presented. The results declared RBF neural networks had an acceptable performance because of having strong academic basic and ability of filtering the noises. This method contain all the experimental data perfectly, and provides the possibility of using different mass fractions and temperatures and simulated charts for knowing the information about the viscosity. For TiO2, ZnO and FeO3 nanoparticles, adding small amounts of nanoparticles decreased the relative viscosity comparing to the base fluid viscosity. But for WO3 and NiO nanoparticles the viscosity of nanofluids was higher than base fluid with any mass fractions.

Learning radial basis function networks with the trust region method for boundary problems

We consider the solution of boundary value problems of mathematical physics with neural networks of a special form, namely radial basis function networks. This approach does not require one to construct a difference grid and allows to obtain an approximate analytic solution at an arbitrary point of the solution domain. We analyze learning algorithms for such networks. We propose an algorithm for learning neural networks based on the method of trust region. The algorithm allows to significantly reduce the learning time of the network.

Solving boundary value problems of mathematical physics using radial basis function networks

A neural network method for solving boundary value problems of mathematical physics is developed. In particular, based on the trust region method, a method for learning radial basis function networks is proposed that significantly reduces the time needed for tuning their parameters. A method for solving coefficient inverse problems that does not require the construction and solution of adjoint problems is proposed.

Optimum steepest descent higher level learning radial basis function network

Dynamically changing real world applications, demands for rapid and accurate machine learning algorithm. In neural network based machine learning algorithms, radial basis function (RBF) network is a simple supervised learning feed forward network. With its simplicity, this network is highly suitable to model and control the nonlinear systems. Existing RBF networks in literature are applied to static applications and also faces challenges such as increased model size, neuron removal, improper center selection etc leading to erroneous output. To overcome the challenges and handle complex real world problems, this paper proposes a new optimum steepest descent based higher level learning radial basis function network (OSDHL-RBFN). The proposed OSDHL-RBFN implements major components inspired from the human brain for efficient learning, adaptive structure and accurate classification. Higher level learning and thinking components of the proposed network are sample deletion, neuron addition, neuron migration, sample navigation and neuroplasticity. These components helps the classifier to think before learning the samples and regulates the learning strategy. The knowledge gained from the trained samples are used by the network to identify the incomplete sample, optimal center and bond strength of hidden & output neurons. Adaptive network structure is employed to minimize classification error. The proposed work also uses optimum steepest descent method for weight parameter update to minimize the sum square error. OSDHL-RBFN is tested and evaluated in both static and dynamic environments on nine benchmark classification (binary and multiclass) problems for balanced, unbalanced, small, large, low dimensional and high dimensional datasets. The overall and class wise efficiency of OSDHL-RBFN is improved when compared to other RBFN’s in the literature. The performance results clearly show that the proposed OSDHL-RBFN reduces the architecture complexity and computation time compared to other RBFN’s. Overall, the proposed OSDHL-RBFN is efficient and suitable for dynamic real world applications in terms of detection time and accuracy. As a case study, OSDHL-RBFN is implemented in real time remote health monitoring application for classifying the various abnormality levels in vital parameters.

Sequential learning radial basis function network for real-time tidal level predictions

Real time tidal level prediction is essential for management of human activities in coastal and marine areas. However, prediction model with static structure cannot represent the variations caused by time-varying factors such as weather condition and river discharge. This paper presents the application of a sequential learning radial basis function (RBF) network for accurate real-time prediction of tidal level. The proposed prediction model employs a sliding data window as dynamical observer, and tunes the structure and parameters of RBF network to adapt to the dynamical changes of tide. The algorithm uses only short-period data as training data and generates predictions sequentially. Hourly tidal data measured at seven tidal stations on west coast of Canada are used to test the effectiveness of the sequential prediction model. Tidal level prediction performance shows that the proposed model can give accurate short-term prediction of tidal levels with very low computational cost.

Performance of auto-configuring RBF networks trained with significant patterns

This paper presents two new ideas to improve the performance of Radial Basis Function (RBF) networks. In the first instance, we choose a set of patterns for training, which are closer to decision boundaries, from different classes of training samples that constitute the input space. We call those set of patterns significant patterns and discuss their selection process from the given data set. Secondly, we use these significant patterns to train Adaptive incremental learning RBF network and Resource Allocating Network (RAN). The learning curves and generalisation characteristics of the generated RBF networks are presented. The performance results are discussed.

Improvement of learning algorithms for RBF neural networks in a helicopter sound identification system

This paper presents a set of optimizations in learning algorithms commonly used for training radial basis function (RBF) neural networks. These optimizations are applied to a RBF neural network used in identifying helicopter types, processing their rotor sounds. The first method uses an optimum learning rate in each iteration of training process. This method increases the speed of learning process and also achieves an absolute stability in network response. Another modification is applied to quick propagation (QP) method as a generalization that attains more learning speed. Finally, we introduced the general optimum steepest descent (GOSD) method, which contains both improvements in learning RBF networks. All modified methods are employed in training a system that recognizes helicopters’ rotor sounds exploiting a RBF neural network. Comparing results of these learning methods with the previous ones yields interesting outcomes.

Real-time performance evaluation of the minimal radial basis function network for identification of time varying nonlinear systems

This paper investigates the real-time performance of the recently developed minimal radial basis function (RBF) network called the minimal resource allocation network (MRAN) on a time varying signal identification problem. MRAN is a sequential learning RBF network and has the ability to grow and prune the hidden neurons to ensure a parsimonious structure. Because of its ability to dynamically adjust its size, MRAN is well suited for real-time signal processing applications. In this paper the ability of MRAN to identify, in real time, a time varying signal is investigated experimentally. In the experiments the sampling rate is varied from low to high values. When the sampling rate increases beyond a limit, the performance of MRAN starts to deteriorate as the time to perform the various calculations in the algorithm begin to exceed the sampling interval. A modification to MRAN that greatly reduces its computational load with very little increase in the overall error is then presented as the real-time alternative to the original algorithm. The real-time MRAN is shown to cope well with smaller sampling intervals for the same time varying signal identification problem.