Differences

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

Link to this comparison view

Both sides previous revision Previous revision
data_mining:svm [2014/08/17 21:05]
phreazer [Parameterwahl]
data_mining:svm [2014/08/17 21:11] (current)
phreazer [Parameterwahl]
Line 95: Line 95:
 $C = \frac{1}{\lambda}$ $C = \frac{1}{\lambda}$
  
-Großes C: Niedriger Bias, hohe Varianz => Overfitting +  * Großes C: Niedriger Bias, hohe Varianz => Overfitting 
-Niedriges C: Hoher Bias, niedrige Varianz => Underfitting+  ​* ​Niedriges C: Hoher Bias, niedrige Varianz => Underfitting
  
 $\sigma^2$ $\sigma^2$
  
-Groß: Features variieren sanfter +  * Groß: Features variieren sanfter => Hoher Bias, niedrige Varianz 
-=> Hoher Bias, niedrige Varianz +  ​* ​Niedrig: Features varrieren abrupter => Niedriger Bias, hohe Varianz
-Niedrig: Features varrieren abrupter +
-=> Niedriger Bias, hohe Varianz+
  
 ===== Kernelwahl ===== ===== Kernelwahl =====