5,337
edits
Machine Learning: Difference between revisions
→VC dimension
| Line 36: | Line 36: | ||
* To show VCdim is at least n, find a training set of size n that can be shattered by our hypothesis class. | * To show VCdim is at least n, find a training set of size n that can be shattered by our hypothesis class. | ||
* To show VCdim is leq n, prove no training set of size n+1 can be shattered by our hypothesis class. | * To show VCdim is leq n, prove no training set of size n+1 can be shattered by our hypothesis class. | ||
* Number of parameters do not necessarily correspond to VC dimension. | |||
: <math>H=\{h(x)=\sin(\theta x)\}</math> has infinite VC dimension with one parameter | |||