Presenting a novel method to improve multi-layered perceptron artificial neural networks based on combination with frog leaping algorithm to detect spam emails

With the advancement of technology and the significant use of the Internet by individuals and groups, advertisements in this field have expanded explosively, so that all kinds of methods for mass distribution of these advertisements to Internet users in the form of e-mail have been encountered. And this causes problems for Internet users. One of these problems is the presence of spam emails which Spam email is one of the most common negative features suffering the owner of an email address. Although existing technologies cannot eliminate unwanted spams, some existing methods can reduce their number. By definition, spam is an electronic version of "useless mails." Spam refers to unwanted and unsolicited e-mails. Such e-mails are not necessarily directly related to the virus, meaning that messages sent from valid sources may also be included in this class. This study proposes a new method based on improving the multilayer perceptron artificial neural network using the shuffled frog leaping algorithm to realize this objective. The shuffled frog leaping algorithm is used to find the best features, and the artificial multilayer perceptron neural network is used to detect spam emails based on the best features. The simulation results prove that the shuffled frog leaping algorithm has significantly improved the multilayer perceptron artificial neural network compared to the artificial neural network based on the radial base function.