Modular granular neural network optimization using the firefly algorithm applied to time series pbkp_rediction

Abstract:

In this chapter the combination of modular neural network, granular computing, and a firefly algorithm is presented to perform time series pbkp_rediction. The Mackey–Glass time series is used to prove the effectiveness of the proposed method. The main contributions of the proposed method are the division of the data points used for the training phase of the modular granular neural network, creating granules of information for each submodule and optimizing the modular granular neural network architecture by a firefly algorithm. Four tests are performed using different numbers of data points for the learning phase and also with and without optimization to prove if there are advantages using the firefly algorithm for neural network optimization.

Año de publicación:

2020

Keywords:

Fuente:

google