Statistical Learning: Concepts

We introduce statistical learning. We distinguish between supervised and unsupervised learning, classification, and regression and between the goals of interpretation and prediction. We introduce the bias-variance tradeoff as a central theme in statistical learning. Next, we introduce the Bayes error as the lowest possible error. The Bayes error is of limited use in practice because it requires knowledge of the true functional relationship between outcome and independent variables. Finally, we reflect on some terms like data science and machine learning. We also explain how language used in statistics differs from that used in computer science.