A Novel Prediction Model using Neural and Fuzzy Temperature Forecasting
Morteza Aghaei, Sina Dami, Seyyed Mohammad Reza Farshchi · 2013
The soft computing techniques especially fuzzy logic has been used by many researchers for temperature prediction in recent years. Artificial neural networks (ANN) have been popular due to their capabilities in handling complex, nonlinear problems in a better way when compared to traditional techniques. In this paper, the theory of ANN with radial basis function (RBF) is presented, and the RBF model is used to predict the daily average temperature for Taipei, Taiwan. The historical data of the daily average temperature and daily cloud density from June 2012 to September 2012 collected from central weather bureau were fed into the RBF model for training and testing. The results were compared with those of previous studies and the performance of the RBF model is found to be prominently better. The expensive experimental result shown, we can conclude that ANN models are more applicable and accurate than fuzzy models to deal with temperature prediction.