Time filter

Source Type

Wu J.,Wuhan University of Technology | Long J.,Guangxi Research Institute of Meteorological Disasters Mitigation | Liu M.,Chengdu University of Technology
Neurocomputing | Year: 2015

In this paper, an effective hybrid optimization strategy by incorporating the adaptive optimization of particle swarm optimization (PSO) into genetic algorithm (GA), namely HPSOGA, is used for determining the parameters of radial basis function neural networks (number of neurons, their respective centers and radii) automatically. While this task depends upon operator's experience with trial and error due to lack of prior knowledge, or based on gradient algorithms which are highly dependent on initial values. In this paper, hybrid evolutionary algorithms are used to automatically build a radial basis function neural networks (RBF-NN) that solves a specified problem, related to rainfall forecasting in this case. In HPSOGA, individuals in a new generation are created through three approaches to improve the global optimization performance, which are elitist strategy, PSO strategy and GA strategy. The upper-half of the best-performing individuals in a population are regarded as elites, whereas the half of the worst-performing individuals are regarded as a swarm. The group constituted by the elites are enhanced by selection, crossover and mutation operation on these enhanced elites. HPSOGA is applied to RBF-NN design for rainfall prediction. The performance of HPSOGA is compared to pure GA in these basis function neural networks design problems, showing that the hybrid strategy is of more effective global exploration ability and to avoid premature convergence. Our findings reveal that the hybrid optimization strategy proposed here may be used as a promising alternative forecasting tool for higher forecasting accuracy and better generalization ability. © 2014 Elsevier B.V.

Huang Y.,Guangxi Research Institute of Meteorological Disasters Mitigation | Jin L.,Guangxi Climate Center
Meteorology and Atmospheric Physics | Year: 2013

A western North Pacific tropical cyclone (TC) intensity prediction scheme has been developed based on climatology and persistence (CLIPER) factors as potential predictors and using genetic neural network (GNN) model. TC samples during June-October spanning 2001-2010 are used for model development. The GNN model input is constructed from potential predictors by employing both a stepwise regression method and an Isometric Mapping (Isomap) algorithm. The Isomap algorithm is capable of finding meaningful low-dimensional architectures hidden in their nonlinear high-dimensional data space and separating the underlying factors. In this scheme, the new developed model, which is termed the GNN-Isomap model, is used for monthly TC intensity prediction at 24- and 48-h lead times. Using identical modeling samples and independent samples, predictions of the GNN-Isomap model are compared with the widely used CLIPER method. By adopting different numbers of nearest neighbors, results of sensitivity experiments show that the mean absolute prediction errors of the independent samples using GNN-Isomap model at 24- and 48-h forecasts are smaller than those using CLIPER method. Positive skills are obtained as compared to the CLIPER method with being above 12 % at 24 h and above 14 % at 48 h. Analyses of the new scheme suggest that the useful linear and nonlinear prediction information of the full pool of potential predictors is excavated in terms of the stepwise regression method and the Isomap algorithm. Moreover, the GNN is built by integrating multiple individual neural networks with the same expected output and network architecture is optimized by an evolutionary genetic algorithm, so the generalization capacity of the GNN-Isomap model is significantly enhanced, indicating a potentially better operational weather prediction. © 2013 Springer-Verlag Wien.

Wu J.,Liuzhou Teacher College | Jin L.,Guangxi Research Institute of Meteorological Disasters Mitigation
Communications in Computer and Information Science | Year: 2011

In this study, a novel hybrid evolutionary algorithm is proposed to improve the regression accuracy of support vector regression (SVR) based on the Particle Swarm Optimization (PSO) and Simulated Annealing (SA) algorithms, called SVR-PSO-SA. This optimization mechanism combined PSO with SA to simultaneously optimize the type of kernel function and the kernel parameter setting of SVR. It is troublesome to escape from the local optima for multi-objective optimization. To avoid premature convergence of PSO, this paper present a new hybrid evolutionary algorithm based on the idea that PSO ensures fast convergence, while SA brings the search out of local optima because of its strong local-search ability. The proposed the hybrid evolutionary algorithm was applied to optimize all parameter setting of SVR for the performance of SVR. The SVR-PSO-SA model is tested at daily rainfall forecasting in Guangxi, China. The results showed that the new SVR-PSO-SA model outperforms the traditional SVR models. Specifically, the new SVR-PSO-SA model can successfully identify the optimal type of kernel function and all the optimal values of the parameters of SVR with the lowest prediction error values in rainfall forecasting. © 2011 Springer-Verlag.

Huang X.-Y.,Guangxi Meteorological Observatory | Jin L.,Guangxi Climate Center | Shi X.-M.,Guangxi Research Institute of Meteorological Disasters Mitigation
Proceedings - 4th International Joint Conference on Computational Sciences and Optimization, CSO 2011 | Year: 2011

Using the capability of extraction the main signal feature from meteorological fields with random noise, and eliminate random disturbance by principal component analysis with conducted on empirical orthogonal functions(EOF), and following the thinking clue of the ensemble prediction in numerical weather prediction (NWP), a novel nonlinear artificial intelligence ensemble prediction (NAIEP) model has been developed based on the multiple neural networks with identical expected output created by using the genetic algorithm (GA) of evolutionary computation. Basing on the sample of typhoon in July from 1980 to 2009 for 30 years in the South China Sea, setting up the genetic neural network (GNN) ensemble prediction (GNNEP) model which selecting the predictors by the method of Stepwise regression and EOF both in the predictors of climatology persistance and Numerical forecasting(NWP) products to predict the typhoon track. The mean error for 24 hours of this new model is 125.7km, and the results of prediction experiments showed that the NAIEP model is obviously more skillful than the climatology and persistence (CLIPER) model with the circumstance of identical predictors and sample cases. © 2011 IEEE.

Jin L.,Guangxi Climate Center | Zhu J.,Center for Ocean Land Atmosphere Studies | Huang Y.,Guangxi Research Institute of Meteorological Disasters Mitigation | Zhao H.-S.,Guangxi Research Institute of Meteorological Disasters Mitigation | And 2 more authors.
Theoretical and Applied Climatology | Year: 2014

Following the practice of the numerical weather ensemble prediction, a nonlinear statistical ensemble prediction model has been developed based on a neural network technique with a Particle Swarm Optimization (PSO) algorithm. The model is validated by short-range climate forecasts of monthly mean rainfall at 37 stations in Guangxi, China during the first rainy season (April, May, and June). Independent prediction results show that the Particle Swarm Optimization Neural Network ensemble prediction model is clearly better than the traditional linear statistical method, such as the multiple regression method and the stepwise regression method. It is also suggested that by applying multiple ensemble members with each member objectively determined by the PSO algorithm, the generalization capacity of the ensemble prediction model is enhanced, demonstrating a vast range of possibilities for operational short-range climate prediction. © 2014, Springer-Verlag Wien.

Discover hidden collaborations