← All skills
Skill · Predicting discrete labelsLive

Classification

Tell classes apart — logistic regression, trees, KNN — and learn the metrics that tell you whether your model is actually any good.

  1. Task 1Your first classifierEasyTrain a model to sort flowers into species.~11 min · 2 blanks→
  2. Task 2Accuracy & the splitEasyScore on unseen data — and keep class balance with stratify.~12 min · 1 blanks→
  3. Task 3K-Nearest NeighboursMediumClassify a point by the company it keeps.~12 min · 1 blanks→
  4. Task 4Decision treesMediumSplit the data into yes/no questions — but not too deep.~12 min · 1 blanks→
  5. Task 5Precision & recallMediumAccuracy lies on hard problems — these two tell the truth.~13 min · 1 blanks→
  6. Task 6The confusion matrixMediumSee exactly which classes get confused for which.~13 min · 1 blanks→
  7. Task 7Probabilities & thresholdsHardGet class probabilities, then decide where to draw the line.~14 min · 1 blanks→
  8. Task 8Multi-classHardTell apart ten classes, not just two.~14 min · 1 blanks→
  9. Task 9Imbalanced dataHardWhen one class is rare, weight it so it isn't ignored.~15 min · 1 blanks→
  10. Task 10Random forestsHardCombine many trees into one strong, stable model.~15 min · 1 blanks→