Optimizing Feedforward Neural Networks Using Biogeography Based Optimization for E-Mail Spam Identification

摘要: Spam e-mail has a significant negative impact on individuals and organizations, is considered as serious waste of resources, time efforts. detection complex challenging task to solve. In literature, researchers practitioners proposed numerous approaches for automatic spam detection. Learning-based filtering one the important used where filter needs be trained extract knowledge that can detect spam. this context, Artificial Neural Networks widely machine learning based filter. paper, we propose use common type Feedforward Network called Multi- Layer Perceptron (MLP) purpose identification, weights network model are found using new nature-inspired metaheuristic algorithm Biogeography Based Optimization (BBO). Experiments results two different datasets show developed MLP by BBO gets high generalization performance compared other optimization methods in literature