Recursive Bayesian echo state network with an adaptive inflation factor for temperature prediction

Biaobing Huang, Guihe Qin, Rui Zhao, Qiong Wu, Alireza Shahriari

Abstract: Temperature prediction is a challenging problem and a concern in energy, environment, industry and agriculture etc. Climate models and statistical time-series forecasting methods are the ineffective forecasting tools of the long-range temperature prediction. A recurrent neural network (RNN) can model complex system with high accuracy. As a type of RNN design approach, echo state network (ESN) is used for temperature forecasting in this study. Based on analysis of monthly maximum, mean and minimum temperatures data sets, a novel recursive Bayesian linear regression (RBLR) algorithm based on ESN is presented in this study. The algorithm consists of two main components: an ESN and a RBLR algorithm with an adaptive inflation factor that changes the confidence level of the prior data. Our proposed method improves the prediction accuracy of the long-range temperature forecasting. Experimental investigations using Central England temperature time series show that the proposed method can forecast monthly maximum, mean and minimum temperatures for the next 12 months and produce good prediction.