Five forecasting algorithms for energy consumption in Vietnam

This paper presents a comparative study of five forecasting models for electric consumption data from Hanoi city, Vietnam. These five forecast models are Adaptive Network Based Fuzzy Inference System (Anfis), Seasonal Auto Regressive Moving Average (Sarima), Multi layer Perceptron (MLP), Elman Recurrent Neural Network (Elman) and Cascade-Correlation Neural Network (CCNN). This study presents an efficient approach in selecting the input variables and the training data structure. The criteria of performance evaluation are computed to estimate and compare these five models. The results indicate in our case a preference for the Multilayer Perceptron, Elman and Cascade-Correlation model. In that the best accuracy prediction is given by the Cascade-Correlation model.