1. Statistical Learning Overview

1-1. Statistics / ML / AI

• data mining: 유한한 표본으로부터 미지의 확률모형을 추정하는 것

1-2. Statistical Learning

Supervised Learning: $P(Y \mid X)$

• goals:
  • prediction
  • inference
  • quantify uncertainty
• learning is not just fitting — generalization under uncertainty

Unsupervised Learning: $P(X)$

• goals:
  • structure estimation
  • clustering
  • representation learning
• models structure in $P(X)$

2. Understanding Data

2-1. Data as Probabilistic Evidence

• data are stochastic evidence
• estimators are random because samples are random!

2-2. Assumptions and Identifiability

• data alone do not uniquely determine the target
• we restrict the set of possible worlds (functional form)
• same data can have multiple explanations