< 8 - Massachusetts Institute of Technology

—1 < < I. can minimize (y— 0 < £(y, h@)) < 4. want to bound h@)) Define £(y,h@)) = From the previous lectures sup — vvrith probabilitv at least 1 — e Define . 9) < Ai I, For now, assume bounds Az 011 sum of 8.veiohtjs (although this IS not true in practice, so we will take union bound later) Theorem 28.1. h(œi)) sup Proof. Since —2 ... ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download