Classification Methods Compared
There is one classifier no method can beat, and it needs the answer to build. Everything else - nearest neighbours, discriminant analysis, logistic regression - is a different guess at what it would have done, and the guesses fail in different directions.
Sign in to take quizzes, earn XP, and unlock stages as you reach 90% mastery.
The Bayes Classifier and Nearest Neighbours
25 min · 100 XPThe rule that minimises test error, the floor it leaves behind, and the nonparametric method that imitates it by counting neighbours - with the number of neighbours turning out to be the flexibility dial in disguise.
Open lesson →Sign in to take the 3-question quiz.
Linear and Quadratic Discriminant Analysis
25 min · 100 XPModel each class’s distribution and turn it around with Bayes’ theorem: one shared covariance gives a linear boundary, one covariance per class gives a quadratic one, and which of them wins is a question about sample size rather than about the truth.
Open lesson →Sign in to take the 3-question quiz.
Thresholds, ROC, and What Accuracy Hides
30 min · 120 XPA single error rate conceals which errors are being made and silently fixes a threshold nobody chose. The confusion matrix separates the two kinds, moving the threshold trades them, and the ROC curve summarises every trade at once.
Open lesson →Sign in to take the 3-question quiz.