The article deals with the fundamental problem of generating a diversity of ensemble predictive models. Various approaches to ensemble diversity are described based on varying the input data set, varying the models used in the ensemble, varying within the model itself, and varying the way the results of the ensemble members are combined. A hybrid method for constructing a heterogeneous ensemble based on various types of neuro-fuzzy networks with a synergistic effect of enhancing the ensemble metamodel is proposed.
Related articles
Other papers on this site that share keywords with this one, closest match first.
International Journal of Distributed Computing and Technology
