Basis Function Network For Prediction Research Articles

Hybrid multiple structural break model for stock price trend prediction

Because of the noises from the internal and external factors, the uncertainty increases in the financial market. The challenges of nonlinearities, volatility clusters, and multiple structural breaks which entail risk. Due to the risk, the prediction task becomes more complex. First, this work proposes a hybrid model to predict the one-day future price for the stocks; MSFT, Apple, Goldman Sachs and JP Morgan use the Markov switching model coupled with radial basis function network for prediction. Second, this paper forecasts the buy/sell trading strategy using the proposed hybrid method. Also, this paper explores the risk of investment decisions and the trading performance based on different value at risk (VaR) methods. Finally, by comparing the proposed model results with the pure linear and non-linear models, the prediction efficiency is evaluated. The experimental results indicate the investment risk, and the investment trading strategy provides a better accuracy with the best investment decision for the selected stocks.

Application of Wilcoxon generalized radial basis function network for prediction of natural gas compressibility factor

Gas compressibility factor is necessary in most of chemical and petroleum engineering calculations. Accurate and fast calculation of this property is of a vital importance in a large number of simulators used in petroleum and gas engineering. In this study, a large data bank (978 data points), covering a wide range of natural gases, was collected from open literature sources. Afterwards, one of the newest and most powerful modeling approach, namely Wilcoxon generalized radial basis function network (WGRBFN) was employed to predict the compressibility factor of natural gases. The results obtained from the proposed model were compared to those of nine empirical correlations and five equations of state. Statistical and graphical error analyses demonstrated that the developed model can satisfactorily predict the compressibility factor of natural gases with an average absolute percent relative error of 2.3%. Moreover, it was demonstrated that the proposed model outperforms all of the studied empirical correlations and equations of state. Finally, to identify the probable outliers the Leverage approach was performed. All of the experimental data seem to be reliable except 2%. Therefore, the developed model is reliable for the prediction of natural gas compressibility factor in its applicability domain.

Determining the structure of a radial basis function network for prediction of nonlinear hydrological time series

A network using radial basis functions (RBFs) as the mapping function in the evolutionary equation for prediction of time series is presented. An RBF network requires the determination of the number of centres of RBFs, their receptive field widths and the linear weights of the network output layer. Traditionally, the number of centres of RBFs is fixed. In this paper, methods to estimate the widths of the receptive fields and the number of centres for the RBFs are introduced. The latter is based on the concept of the generalized degrees of freedom. The linear weights are determined by the least-squares method. The proposed method is then applied to make predictions of six sets of data: two theoretical functions that are known to become chaotic under certain parameter conditions (Henon map and Lorenz map), and four real-time series (discharge data from the Mekong River in Thailand and Laos, and from the Chao Phraya River in Thailand, and sea-surface temperature anomaly data). The results are at least one order of magnitude better than those obtained by a similar model with fixed number of centres as well as by a linear model and a stochastic model for most of the data sets.