data_mining:neural_network:short_overview

Differences

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

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
data_mining:neural_network:short_overview [2018/04/22 00:16] – [Calculation of predictions errors] phreazerdata_mining:neural_network:short_overview [2018/05/10 17:32] (current) – [Cost function] phreazer
Line 96: Line 96:
   * Sum over k: number of outputs   * Sum over k: number of outputs
   * Sum over all $\theta_{ji}^{(l)}$ without bias units.   * Sum over all $\theta_{ji}^{(l)}$ without bias units.
 +  * Frobenius norm for regularization, also called //weight decay//
  
 ===== Backpropagation Algorithm ===== ===== Backpropagation Algorithm =====
Line 108: Line 109:
 a^{(1)} = x \\ a^{(1)} = x \\
 z^{(2)} = \theta^{(1)} a^{(1)} \\ z^{(2)} = \theta^{(1)} a^{(1)} \\
-a^{(2)} = g(z^{(2)}) \text{ füge } a_0^{(2)} \text{hinzu} \\+a^{(2)} = g(z^{(2)}) \text{ add } a_0^{(2)} \\
 \dots \dots
 $$ $$
Line 132: Line 133:
 Algorithmus Algorithmus
  
-$$\Delta_{ij}^{(l)} = 0 \text{für alle i,j,l} \\+$$\text{Set } \Delta_{ij}^{(l)} = 0 \text{ for all i,j,l} \\
 \text{For i=1 to m:} \\  \text{For i=1 to m:} \\ 
-\text{Set} a^{(1)} = x^{(i)}$$+\text{Set } a^{(1)} = x^{(i)}$$
  
 Forward propagation to compute $a^{(l)}$ für $l=2,3,\dots,L$ Forward propagation to compute $a^{(l)}$ für $l=2,3,\dots,L$
  • data_mining/neural_network/short_overview.1524349016.txt.gz
  • Last modified: 2018/04/22 00:16
  • by phreazer