K-Nearest Neighbor.md

1-NN

![[Pasted image 20240530144723.png]]

K-NN

距离度量

![[Pasted image 20240530144933.png]] ![[Pasted image 20240530144943.png]]

属性

属性归一化

![[Pasted image 20240530145333.png]]

属性加权

![[Pasted image 20240530145348.png]]

连续取值目标函数

![[Pasted image 20240530165425.png]]

$k$ 的选择

![[Pasted image 20240530165519.png]]

打破平局

![[Pasted image 20240530165627.png]]

关于效率

![[Pasted image 20240530165728.png]] [[KD-Tree]]

优点

![[Pasted image 20240530170329.png]] ![[Pasted image 20240530170341.png]]

下一个问题

![[Pasted image 20240530170533.png]]

Distance-weighted KNN

![[Pasted image 20240530170716.png]]

基于实例的学习器的四个要素总结

![[Pasted image 20240530170942.png]] ![[Pasted image 20240530172513.png]]![[Pasted image 20240530172519.png]] ![[Pasted image 20240530172530.png]]