Micro-F1

With more than two classes there's no single "positive" class, so precision and recall need pooling. Micro-averaging pools the counts themselves: total up true positives, false positives and false negatives across every class, then compute one precision, one recall, one F1.

For each class c, counting over all samples:

• TP — predicted c, actually c
• FP — predicted c, actually something else
• FN — predicted something else, actually c

Then precision = TP / (TP + FP), recall = TP / (TP + FN), and F1 = 2·P·R / (P + R).

Task: write micro_f1(y_true, y_pred) returning the micro-averaged F1, rounded to 4 decimal places.

• Consider every class appearing in either list.
• Return 0.0 if there are no true positives at all.

Micro-averaging is dominated by whichever classes have the most samples — a rare class barely moves it. That's the opposite of macro-averaging, which gives every class an equal vote regardless of size. Neither is more correct; they answer different questions.