Naive Bayes spam filter guided project

Hi,

I completed the guided project for building a Naive Bayes SMS spam filter. I was looking through the solutions and see that the solution managed to achieve a 98.74% accuracy, where I only managed to achieve 86.6%.

My solution
DQ solution

I even adjusted some of my solution to the one provided to improve accuracy, but cannot seem to get it up. Can anyone suggest a reason for this 12% difference in accuracy between solutions?

Thank you

1 Like