Boosting Sketch, Ranking, and Voting

Status: placeholder — survey depth only.

Goals¶

• Weak vs strong learning (idea)

• Boosting as a sketch (AdaBoost intuition)

• Ranking: Borda count, comparison to matrix methods

• Voting: basic rules and paradoxes

• Optional footnote: PageRank as linear algebra (not a Markov-chain chapter)

Outline¶

1. Combining weak classifiers

2. Boosting idea (no full proofs)

3. Aggregating preferences: ranking & voting

4. Student-talk-friendly topics from this page

Sources¶

• Blum, Hopcroft, Kannan, Foundations of Data Science, selected sections