Hybrid Model for Email Spam Detection Using Multi-Modal KNN and SVM

Email spam is one of the big problems and really accelerates online fraud against our society today. Phishing and fraud became one of the nefarious and illegal uses by some people on this. The creation of a fake account and profile gives an opportunity to spammers to sound more valid in the emails they send, thus making it much easier for spammers to spam. This paper applies the multi-modal based K-Nearest Neighbour algorithm (KNN) and support vector machine (SVM) learning techniques for presenting an approach toward email fraud detection. This technique is known to be straightforward and very efficient while working with huge datasets. It can be applied to model the correlations existing between the words in the emails and their characteristics as spam or ham. SVM is able to separate classes in high dimensional spaces and also used for accuracy classification. This paper assumes the integration of both methods to propose a method aiming to enhance the detection of spam emails based on the advantages, both of the algorithms may provide. This mixed-breed model, in this current research, is presented as an incredibly potent solution to this problem, so that spam and not spam emails are distinguished from one another. Most of the approaches in present methods are found to have the lesser accuracy in spam detection. With the proposed Multi-Modal KNN and SVM algorithm, the accuracy level reached is 99.23%, resulting in faster processing speed as compared to other proposed methods.