Zu dieser Karteikarte gibt es einen kompletten Satz an Karteikarten. Kostenlos!

k-NN works best when there is lots of data avaiable and the data has a small amount of features.

Pros:

• Easy to implement
• Very fast training
• No information loss
• high classification accuracy if lots of data is avaiable
• Intuitive interpretation
• Can have very complex decision boundaries

Cons:

• Requires lots of memory to store all the data samples
• Slow query time
• Sensitive to the local structure of the data
• The parameter k needs to be tuned