Using Machine Learning Algorithm to Enhance Phishing Emails Classification

Phishing emails are becoming a more hazardous problem in online banking and social networking platforms like Facebook, Twitter, and Instagram. Phishing is usually done by mimicking an email or embedding material in the email body, prompting the recipient to give their credentials. Users forget their training tactics and warning signals after a while, therefore phishing approach training is ineffective. It is completely reliant on the user’s actions, which will be recorded at a certain time in reaction to software warning messages while accessing any URL. Using the Spam basic dataset as a starting point, This paper uses J48, Nave Bayes, and a decision tree to improve phishing email classification. With a true positive rate of 97 percent and a false negative rate of 0.025 percent, J48 is the best spam categorization. Random forest works best with small datasets of up to 5000 features and 34 features. When the dataset size and number of features are lowered, however, Nave Bayes performs better.

Author (S) Details

Vidya Mhaske-Dhamdhere
Bharati Vidyapeeth (Deemed to be University) College of Engineering, Pune, India.

Sandeep Vanjale
Bharati Vidyapeeth (Deemed to be University) College of Engineering, Pune, India.

View Book :-  https://stm.bookpi.org/CTMCS-V4/article/view/2285

Leave a Reply

Your email address will not be published. Required fields are marked *