Python Idle
A model with 85% accuracy on spam detection sounds great — until you realize just predicting "ham" every time gets you 85%. Your spam filter is basically a rubber stamp. Time to look deeper at what your model actually gets right and wrong, and arm yourself with metrics that tell the real story.
~15 minscenario5572 rows
Loading Python runtime...
Goals: 3 tests
should prove dummy accuracy is misleadingly high (>80%)
should train a real model with higher F1 than the dummy
should calculate precision and recall
Python loading...