“…Up to now, several hybrid prediction models have been proposed such as combined ARIMA-support vector machine (SVM) [20][21][22], hybrid of Grey and Box-Jenkins autoregressive moving average (ARMA) models [23], hybrid of ARIMA and fuzzy logic [24], hybrid of support vector regression (SVR) and differential evolution (DE) algorithm [25], integrated ANN-genetic algorithms (GAs) [26][27][28]/ANN-particle swarm optimization (PSO) [14]/ ANN-artificial fish swarm algorithm (AFSA) [29], combined generalized linear autoregression (GLAR)-ANN [30], hybrid of artificial intelligence (AI) and ANN [31], hybrid of wavelets and ANN implemented on a decision support system [32,33], integrated ARMA-ANN [34,35], combined seasonal ARIMA-back propagation (BP) ANN [36], combination of several ANNs [15,37], hybrid model of self organization map (SOM) neural network, GAs, and fuzzy rule base (FRB) [38], combined fuzzy techniques-ANN [39][40][41], hybrid based on PSO, evolutionary algorithm (EA) and DE for training a recurrent neural network (RNN) [42].…”