Zusammenfassung der Ressource
Sampling
- Simple Random Sample
- Seed number. The chance of selection is equal
- Elements selected independently
- Systematic
- Taking every nth observation
- Take the first n examples
- Stratified - Classifying
the data, and taking
random samples
- Unbalanced stratified
- Taking a random sample, from each class not proportional to the class sizes
- Overfitting
- Perfect fits to the data,
when its like in reality unlike
to be perfect