Differences

This shows you the differences between two versions of the page.

--- data_mining:large_datasets [2014/09/06 14:38] – [Stochastic gradient descent] phreazer
+++ data_mining:large_datasets [2014/09/06 14:45] (current) – [Online Learning] phreazer
@@ Line 50: / Line 50: @@
 Verwende b = mini-batch size (typisch: 2-100)  Beispiele in jeder Iteration.
+===== Online Learning =====
+Wenn kontinuierlicher Datenzufluss.
+Lernen von $p(y=1|x;\theta)$ um Preis zu optimieren.
+  Repeat forever {
+    Get (x,y) corresponding to user.
+    Update \theta using (x,y)
+      \theta_j = \theta_j - \alpha (h_\theta(x)-y) x_j
+  }