Review Note
Last Update: 10/08/2023 02:57 PM
Current Deck: EEN100: Theory and exercises
New Card (Unpublished)Currently Published Content
Front
Back
No published tags.
Pending Suggestions
Field Change Suggestions:
Front
Commit #21167
\(\epsilon\)-representative sample
Back
Commit #21167
A training set \(\mathcal S\) is called \(\epsilon\)-representative if for all \(h \in \mathcal H\) it holds that \[\lvert L_{P_{\mathcal Z}}(h) - L_{\mathcal S}(h)\rvert \le \epsilon\]i.e. training error and population error are \(\epsilon\) close for any prediction rule.