We'll use an Artifical Neural Network to classify messages as "spam" or "no spam".
The goal is to clean up this mess:
The Problem
β...
For further actions, you may consider blocking this person and/or reporting abuse
Looks great! Bookmarked for careful reading later.
Why has it very little to do with ML? Naive Bayes algorithms (Gaussian, Bernoulli etc.) are used like any other supervised ML classifier, no?
In regards to comparing NB & ANN, I still think it would be interesting to test it, e.g. in this project. I will take a look at how NB performs here.
Additionally, I only found this comparative study: pdfs.semanticscholar.org/d945/29ae...
Would be great to see more NB vs ANN comparisons, though.
Good point. What I left out in this project is comparing the performance of differently ML solutions, like naiΜve Bayes, decision trees, random forests etc.
Training these classifiers on our problem and evaluating and comparing the F1 score would actually be a good benchmark to compare our ANN against.