Efficiency of local models ensembles for time series prediction

► Two neural-based local modelling approaches for prediction are proposed. ► Low computacional time and scalability to data set size, are stressed. ► Performance comparison with other global and local approaches is showed.