A Hybrid Material Generation Algorithm with Probabilistic Neural Networks for Solving Classification Problems

• DOI: 10.14569/ijacsa.2022.0130532
• Publication Date: 2022-06-01T05:46:32Z
• Authors: Mohammad Wedyan; Omar Alshaweesh; Enas Ramadan; Ryan Alturki; Foziah Gazzawe; Mohammed J. Alghamdi
• Keywords: Firefly Algorithm; Benchmark (surveying); Data classification; Feedforward neural network

Classification is based on machine learning, in which each element a set of data classified into one predetermined groups. In mining, an artificial neural network (ANN) the most significant methodology because exact results obtained through this algorithm and applied solving many classification problems. ANN consists group types feed-forward networks, feed-back network, RFB probabilistic networks (PNN). For issues, PNN frequently utilized. The primary goals research are to fine-tune weights enhance classification accuracy. To accomplish goal, Material Generation Algorithm (MGA) was investigated with hybrid model. Newly, hybridization algorithms ubiquitous it has led development unique procedures that outperform those use single algorithm. Several distinct tasks used test efficiency suggested (MGA-PNN) approach. MGA algorithm's evaluated using training outcomes generated, its compared other optimization strategies. By 11 benchmark datasets, performance terms accuracy evaluated. display outperforms biogeography optimization, firefly method