diegoocampoh / machinelearningphishing Goto Github PK

The results section states that:
Currently the algorithm that performs bests in this scenario and with this email corpus is Naive Bayes with an accuracy of 99,18% and a recall of 98,46%. These results may vary when the project is done, as value normalization and other optimizations are due.

But the chart included in that section lists naive Bates at 61.5% accuracy and recall of 22.8%.

I want to input my custom emails and convert it to dataset format , so that i can check whether my mail is phishy/not.