Comparing performances of backpropagation and genetic algorithms in the data classification

Artificial neural networks (ANN) have a wide ranging usage area in the data classification problems. Backpropagation algorithm is classical technique used in the training of the artificial neural networks. Since this algorithm has many disadvantages, the training of the neural networks has been implemented with the binary and real-coded genetic algorithms. These algorithms can be used for the solutions of the classification problems. The real-coded genetic algorithm has been compared with other training methods in the few works. It is known that the comparison of the approaches is as important as proposing a new classification approach. For this reason, in this study, a large-scale comparison of performances of the neural network training methods is examined on the data classification datasets. The experimental comparison contains different real classification data taken from the literature and a simulation study. A comparative analysis on the real data sets and simulation data shows that the real-coded genetic algorithm may offer efficient alternative to traditional training methods for the classification problem.

Research highlights
► Data classification is one of the fundamental problems in many decision-making tasks. Artificial neural networks (ANN) have popularity in the data classification problems. One of the important issues on the neural networks is training of the networks. Backpropagation algorithm is the most widely used search technique for training neural networks. Backpropagation algorithm has negative properties such as being captured in the local solutions and having low classification performance in some cases. In order to prevent these disadvantages, researchers have proposed many alternatives. Genetic algorithms are efficient alternatives for training of the neural networks. It is known that the comparison of the approaches is as important as proposing a new classification approach. In this study, the training of the ANNs for the classification problems is examined by the backpropagation, binary-coded and real-coded genetic algorithm. In order to compare these training algorithms, 10 different real-world and large-scale simulation datasets covering the different network architectures is used. The results based on real-data and simulation show that classification success of artificial neural network model trained with real-coded genetic algorithm is better than other training methods.