Application of artificial intelligent models for prediction of daily air temperature in Rasht station

Estimation of air temperature is importan in environmental and agricultural sciences. The aim of this study is prediction of daily air temperature (mean, maximum and minimum) using several types of AI models, in Rasht station north of Iran. According to the results, the adaptive neural-fuzzy inference system, artificial neural network, and gene expression programming were ranked first, second and third, respectively, despite of a slight difference in prediction accuracy of the selected models. Besides, the developed mathematical equation between the input and output variables using the gene expression programming model showed the superiority of this approach to the other two models. Based on the SI index for the minimum and mean daily temperature in training period is varied in range of 0.1 to 0.2, i.e acceptable in terms of model accuracy. For the maximum temperature it ranged 0.2 to 0.3 which is considered as average accuracy. The findings revelead the best performance can be obtained using inputs in one to three days lead time.