Radial Basis Function Network Training Using a Nonsymmetric Partition of the Input Space and Particle Swarm Optimization

Alexandridis, Alex; Chondrodima, Eva; Sarimveis, Haralambos

doi:10.1109/tnnls.2012.2227794

Cited by 112 publications

(61 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, there are abundant of proposals developed with their own merits and demerits such as constructive decay [29], resource allocating networks [55], and the minimum description length principle [56]. Recently, Alexandridis et al, [57], introduced a nonsymmetric approach for partitioning the input space. Their experimental outcomes have shown that the nonsymmetric partition can lead to the development of more accurate RBF models, with a smaller number of hidden layer nodes.…”

Section: Orthogonal Least Squares (Ols)mentioning

confidence: 99%

Radial basis function neural networks: a topical state-of-the-art survey

Dash¹,

Behera²,

Dehuri

et al. 2016

Open Computer Science

View full text Add to dashboard Cite

Radial basis function networks (RBFNs) have gained widespread appeal amongst researchers and have shown good performance in a variety of application domains. They have potential for hybridization and demonstrate some interesting emergent behaviors. This paper aims to o er a compendious and sensible survey on RBF networks. The advantages they o er, such as fast training and global approximation capability with local responses, are attracting many researchers to use them in diversi ed elds. The overall algorithmic development of RBF networks by giving special focus on their learning methods, novel kernels, and ne tuning of kernel parameters have been discussed. In addition, we have considered the recent research work on optimization of multi-criterions in RBF networks and a range of indicative application areas along with some open source RBFN tools.

show abstract

Section: Orthogonal Least Squares (Ols)mentioning

confidence: 99%

Radial basis function neural networks: a topical state-of-the-art survey

Dash¹,

Behera²,

Dehuri

et al. 2016

Open Computer Science

View full text Add to dashboard Cite

show abstract

“…Radial basis function (RBF) neural networks due to their simple architecture and computational efficiency have attracted the attention of many researchers [19][20][21][22], and particle swarm optimization (PSO) is a global optimization strategy for solving continuous and discrete optimization problems [23][24][25][26][27]. The hybrid learning algorithm of the PSO-based RBF neural network has also been used for many applications [28][29][30][31]. In this paper, the global optimum parameter of the RBF is determined by means of PSO, thus eliminating the chattering or steadystate error happened in RNFCSMC.…”

Section: Introductionmentioning

confidence: 99%

Hydrogen Energy paper 13

2018

Proceeding of 2018 the Fourth International Symposium on Hydrogen Energy, Renewable Energy and Materials

View full text Add to dashboard Cite

Abstract. In this paper, a robust nonsingular fast converging sliding mode control (RNFCSMC) with particle swarm optimization (PSO)-based radial basis function (RBF) neural network is presented and applied in hydrogen fuel cell systems capable to maintain low harmonic distortion even in case of nonlinear load. The proposed scheme is a modified structure: a RNFCSMC plus a PSO-based RBF neural network. Though the classic sliding mode control has inherent robustness against plant parameter variations and load disturbances, the convergence of the system states to the zero is usually asymptotical in infinite time. The RNFCSMC is introduced to assure the finite time convergence of the system states and there is no singularity problem. But, once a severe load disturbance is applied, the chattering or steady-state error still exists in. The PSO-based RBF neural network is employed to determine the control gains of the RNFCSMC, thus eliminating the chattering or steady-state error so that the system performance reaches the optimal point. The proposed technical method has been realized for the actual singlephase hydrogen fuel cell inverters controlled by a DSP, and the simulations and experiments clearly demonstrate the effectiveness of the proposed technical method. Because the proposed hydrogen fuel cell system offers significant advantages over classic sliding mode system in terms of complexity and ease of implementation, this paper will be a useful reference to related control designers of hydrogen fuel cell systems.Keywords: robust nonsingular fast converging sliding mode control (RNFCSMC), particle swarm optimization (PSO)-based radial basis function (RBF) neural network, hydrogen fuel cell systems, singularity problem, chattering.

show abstract

“…The search rules implemented by the PSO algorithm are simpler than those encountered in the GA and the AIA, and PSO algorithm is easy to implement and exhibits quick convergence. In light of these advantages, recently the PSO algorithm has attracted a lot of attention as an effective optimization tool [11,21,23]. (3) Given the kernel matrix problem present in SVR, some researchers employ data clustering and data pruning methods to reduce computing time, such as K-Means clustering [12], Fuzzy C-Means [13], and clustering kernel row vectors [24].…”

Section: Introductionmentioning

confidence: 99%

“…Many researchers have pointed out that three crucial problems existing in SVR urgently need to be addressed: (1) How to choose or construct an appropriate kernel to complete forecasting problems [8,9]; (2) How to optimize parameters of SVR to improve the quality of prediction [10,11]; (3) How to construct a fast algorithm to operate in presence of large datasets [12,13]. With unsuitable kernel functions or hyperparameter settings, SVR may lead to poor prediction results.…”

Section: Introductionmentioning

confidence: 99%

Global Nonlinear Kernel Prediction for Large Data Set With a Particle Swarm-Optimized Interval Support Vector Regression

Ding

Cheng

Pedrycz

et al. 2015

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Abstract-A new global nonlinear predictor with a particle swarm optimized interval support vector regression (PSO-ISVR) is proposed to address three issues (viz. kernel selection, model optimization, kernel method speed) encountered when applying support vector regression (SVR) in the presence of large datasets. The novel prediction model can reduce the SVR computing overhead by dividing input space and adaptively selecting the optimized kernel functions to obtain optimal SVR parameter by PSO. To quantify the quality of the predictor, its generalization performance and execution speed are investigated based on statistical learning theory. In addition, experiments using synthetic data as well as the stock volume weighted average price (VWAP) are reported to demonstrate the effectiveness of the developed models. The experimental results show that the proposed PSO-ISVR predictor can improve the computational efficiency and the overall prediction accuracy compared with the results produced by the SVR and other regression methods. The proposed PSO-ISVR provides an important tool for nonlinear regression analysis of big data.Index Terms-global nonlinear predictor, interval support vector regression, particle swarm optimization, kernel function, sliding adaptive model, large data I. INTRODUCTIONUPPORT vector regression (SVR) model is constructed based on statistical learning theory [1], which uses a kernel function to map the data from some input space to a highdimensional feature space where the problem becomes amenable for handling by linear regression [2]. Owing to its robustness to noise and its generalization abilities, it has been widely employed in various areas such as adaptive flight identification [3], ore grade estimation [4] , and stock market price forecasting [5][6][7].Many researchers have pointed out that three crucial problems existing in SVR urgently need to be addressed: (1) How to choose or construct an appropriate kernel to complete forecasting problems [8,9]; (2) How to optimize parameters of SVR to improve the quality of prediction [10,11]; (3) How to construct a fast algorithm to operate in presence of large datasets [12,13]. With unsuitable kernel functions or hyperparameter settings, SVR may lead to poor prediction results. In fact, a kernel function forms a certain nonlinear transformation function. Due to data uncertainty in practical regression problems, it is difficult to determine which kernel function is the best one for a specific problem without any prior knowledge [14]. If the adjustable kernel parameters in SVR are not properly selected, it will result in the undesirable phenomena of over-fitting or under-fitting [1]. Furthermore, SVR is typically confronted with a heavy computing overhead due to processing large Gram matrices being associated with the kernels [13]. This computing burden becomes an essential barrier when dealing with massive data, such as those encountered in protein structure prediction and time series prediction [15].During the past several years, various methods h...

show abstract

Radial Basis Function Network Training Using a Nonsymmetric Partition of the Input Space and Particle Swarm Optimization

Cited by 112 publications

References 52 publications

Radial basis function neural networks: a topical state-of-the-art survey

Radial basis function neural networks: a topical state-of-the-art survey

Hydrogen Energy paper 13

Global Nonlinear Kernel Prediction for Large Data Set With a Particle Swarm-Optimized Interval Support Vector Regression

Contact Info

Product

Resources

About