Abstract— Objective: The aim of this paper is to analyze the development of new forecasting models based on neural networks using the guidelines of the synthesis method, systematic literature review. Method: A systematic literature review method with a manual search of papers published in the last 15 years (2000 to 2014) on new neural networks models for time series forecasting is presented. Results: Only 19 studies meet all the requirements of the inclusion criteria. Of these, only three proposals considered a neural networks model using a process different to the Autoregressive. Conclusion: Although studies relating to the application of neural network models were frequently present, with few studies that proposing new neural networks models for forecasting with theoretical support systematic procedure for the construction of model in the period of study.
Keywords — Nonlinear models, Neural networks models, Forecasts, Innovation, Search methods.