Study	Study methodology	Our methodology
Fette et al. [5]	They used two non-spam, non-phishing datasets with 10 features on their proposed approach the overall accuracy they have achieved was 99.5	We used three datasets with different features , our result more accurate since they’ve used an imbalanced dataset
Bhat et al. [6]	They used a text dataset consists of 1897 spam, 3900 ham, and 250 ‘hard’ ham The result identified Random Forests as the classifier with 98.3 accuracy	We used in the third experiment text dataset with 2500 ham emails and 500 spam emails, In our experiments we used seven algorithms, highest accuracy that we got is Neural Network algorithm with 97.7 accuracy
Islam et al. [8]	They used public data sets PUA and used three classification algorithms as NB, SVM and AdaBoost, It has been shown that the accuracy of their proposed system (97.05)	We used three datasets with different features and seven algorithms the best ML algorithm accuracy rates achieved was for boosted decision tree and Neural Network