Karteikarten
Machine learning questions
Karten in diesem Stapel (5)
- 1What decides an appropriate split position when partitioning a feature space?- In the case of linear regression: we may use residual sum of squares (RSS) to see how good predictions are
- In the case of classification trees, we use impurity measures - 2What is the formula for RSS?
- 3What are impurity measures?Criteria which contrast the observed section of observations from a certain class to those predicted by the model.
- 4When will a node be impure?If you classify a certain partition to belong to a certain class, but many observations that can be found in that partition do not belong to that class, the node is said to be impure.
- 5How do we measure impurity?There are three different methods:
Misclassification error
Shannon-entropy
Gini-index
Misclassification rate is the least sensitive but Shannon-Entropy is the most sensitive.
Informationen
- Beschreibung
- Statistics Karteikarten am Machine learning questions, erstellt von Ciara Madella am 03/10/2019.
- Erstellt von
- Ciara Madella · Okt 2019
- Aktualisiert
- 3 Okt 2019
- Aufrufe
- 0
- Bewertung