A Fuzzy Neural Network Precipitation Model Based on Rough Set
Yi Qing Zhu · Jisuanji fangzhen · 2009
To improve the predictive ability of a fuzzy neural network(FNN) prediction model,the paper presents a prediction modeling methodology which combines the rough set attribute reduction with FNN.Taking short-term precipitation prediction as research objects and using attribute reduction to compute and analyse the original selected matrix of forecast factors,the scale of network is effectively reduced and the prediction ability of the FNN prediction model is evidently enhanced.Result shows that the FNN model based on rough set is superior to the conventional regression and FNN prediction model based on regression and CMA T213 numerical prediction model.Eventually,the merits of the rough set attribute reduction and FNN techniques are explained.