Detecting Spear-phishing Emails Based on Authentication

Phishing Mitigation Techniques: A Literature Survey

“…Subsequent emails are compared to these models which are developed using a Support Vector Machine for classification. [2]improved it by combining stylometric features, gender features, and personality features.Their approach uses feature extraction to build and keep an identity profile model of a sender; hence subsequent emails of the sender are compared against the profile, and in a case where the profile of an uncertain email is consistent with the legitimate profile of the sender, the sender of the uncertain email is identified as legitimate and the email is considered normal mail. However, if the profile of an uncertain email is inconsistent with the legitimate profile of the sender, the sender is masked, and the email is classified asaspear-phishing email.…”

“…The rule-based approaches are among the earliest solutions which are proposed for spam detection [47].This approach performs well on known set rules. However, it has high false alarm rates and its difficulty in updating rules in case of big data is also a challenge to this approach [2].Therefore, a research in the area of rule-based to reduce the time it takes to update rules would definitely help in combat phishing attack as black and whitelists have proven to be effective approaches to phishing attacks but the manual updating did not help it, hence a need for automatic rule update system in the organizations for the safety of the users.…”

“…In general, the rule-based approaches perform well on known set rules. However, it has high false alarm rates and its difficulty in updating rules in case of big data is also a challenge to this approach [2]. The updating issue with rule-based approaches if solved would reduce if not eliminated the falsealarm rate of this approach and the performance accuracy would as well increase.…”

Online Social Networks and Writing Styles–A Review of the Multidisciplinary Literature

“…Authorship verification can also be applied for detecting phishing emails. [113] outperformed the two common phishing email detection methods, PILFER and FSSPD, by 10%. The researchers combine stylometry (97 features), gender features (7 features), and personality features based on emotion words (15 features) in order to verify the author of a phishing attack, which is also known as spear-phishing.…”

“…In the cybersecurity and personality & behavior disciplines, researchers are seen to use keywords that represent demographics such as gender and age as part of the features to identify the author. There are also some papers in the cybersecurity discipline that either use personality features or combine both demographics and personality features into their detection methods [111], [113], [117], [132]. Therefore, future studies regarding writing styles will also benefit by combining multidisciplinary domains as people are complex and not one dimensional.…”