An architectural-based approach to detecting spim in electronic means of communication

Abstract

Spams are what users and developers should be aware of in all Internet-based communication tools (such as e-mail, websites, Social Networking Sites (SNS), instant messengers and so on). This is because spammers have not ceased from using these platforms to deceive and lure users into releasing vibrant and sensitive information (especially, financial details). This paper developed an architectural based technique for SPIM (Instant Message Spam or IM SPAM) detection using the classification method. The classification was done using the C4.5 classifier with a dataset of messages gotten from an instant messaging environment. The dataset served as the input to the classification algorithm method which was able to distinguish spam from non-spam messages. This classification method was depicted in a tree form to show its usefulness. The results show that its precision, recall and accuracy rate satisfied standard recommendation with a commendable error rate. The proposed technique will find implication in the reduction of the number of Internet users.

Filtering Spam Messages Based on Improved Naive Bayesian Algorithm

“…After removing the stop words, it is necessary to select the feature words to lay a solid foundation for the subsequent feature expansion. For feature selection, an improved TF-IDF algorithm is adopted, which can extract key feature words well and focuses on the distribution of feature words among different categories in a category [18].…”