Accuracy
ELI5 — The Vibe Check
Accuracy is the simplest way to score a model — what percentage of predictions were correct. 95% accuracy sounds great, until you realize 95% of your emails are not spam, so a model that never flags anything gets 95% accuracy too. Accuracy lies. Always check precision and recall too.
Real Talk
Accuracy is the fraction of correct predictions out of total predictions: (TP + TN) / (TP + TN + FP + FN). It is a useful metric when classes are balanced but misleading for imbalanced datasets. A spam filter with 99% non-spam data can achieve 99% accuracy by predicting 'not spam' for everything.
When You'll Hear This
"The model hit 98% accuracy on the test set." / "Accuracy is misleading on this imbalanced dataset."
Related Terms
Classification
Classification is teaching an AI to sort things into categories. Is this email spam or not? Is this image a cat, dog, or bird?
F1 Score
The F1 Score is the balanced average of precision and recall — a single number that captures both.
Overfitting
Overfitting is when your model gets TOO good at the training data and becomes useless on new data.
Precision
Precision asks: 'Of all the times the AI said YES, how often was it actually right?
Recall
Recall asks: 'Of all the actual YES cases in the world, how many did the AI catch?' High recall means the model finds almost everything it should.