Detecting Targeted Malicious Email Through Mail Client

Abstract

Sharing and storing of data in the web world is with the help of social networks. Messages are exchanged between hosts using the Simple Mail Transfer Protocol (SMTP). While the email messages are transporting between systems, SMTP communicates delivery parameters using a message envelope separate from the message (header and body) itself. A malicious email message is the one which have been deliberately crafted to cause problems on the server or at the client side. This message may contain a virus. A filtering technique is applied on messaging string contents by applying tokenization and then applying naïve bayesian classifier we classify targeted and non targeted malicious email. A network defender encounters different classes of threat actors with varying intents and capabilities. Conventional computer network attacks exploit network-based listening services such as Web servers. Traditional decision-tree classification algorithms split each node using the best split from all available features. With random forests, each node splits from a randomly selected set of features at that node. It is only focus on mail body not on hyperlinks and attachments. Proposed extension is feature extraction to file attachment metadata. Threat actors might inadvertently leave remnants of infor-mation such as file paths, time zones, or even author names. In addi¬tion, organizations can track features that characterize the types and amounts of email received by a particular email address.