Performance Analysis of Artificial Neural Network - Chimpanzee Leader Election Optimization on Classification Case
Ferry Wahyu Wibowo, Akhmad Dahlan, Wihayati
Abstract
Artificial neural network (ANN) modeling has a long history of modeling and long-standing model modifications. In addition, various ways researchers and programmers do improve ANN performance. Although many developments have been carried out on ANN, this modeling continues to be explored further. This paper aims to combine ANN modeling and one of the population-based metaheuristic models, namely the chimpanzee leader election optimization (CLEO) algorithm. The CLEO algorithm is an optimization algorithm that models male chimpanzees as the leader in their community. The CLEO algorithm has the ability that it does not use hyper tuning in its application. It is because the CLEO method only relies on random values on its variables. The CLEO method in hybrid modeling functions in finding solutions to the ANN weight values. This study observes the performance of ANN-CLEO for classification tasks. The observation results show that the best classification accuracy for the results of this hybrid method depends on the domain of finding the solution or the given upper and lower limit values.