(1) Euclidean distance measures(欧氏距离) Euclidean distance is the most commonly used measure Euclidean distance between observations I and r d x. I; =(x1-x1)+(X2-x2)'++(xi-x i) 2021/1/21
2021/1/21 21 cxt (1)Euclidean Distance Measures(欧氏距离) Euclidean distance is the most commonly used measure Euclidean distance between observations and
o since the square root operation does not change the order of how close the observations are to each other. some programs use squared Euclidean distance (i.e. they do not take the square root) ∑(x-x)=(x :(x1-xn)2+(x2-X2)2+…+(x-X k=1 2021/1/21
2021/1/21 22 cxt ❖ Since the square root operation does not change the order of how close the observations are to each other, some programs use squared Euclidean distance (i.e., they do not take the square root). 2 2 2 2 2 1 1 2 2 1 ( ) ( ) ( ) ( ) p ij ik jk i j i j ip jp k d x x x x x x x x = = − = − + − + + −
(2) Minkowski metric(明考夫斯基距离) d xi, X 0-3kxa-xp 2 When m=1, the measure 21xx-xAs called the city-block or Manhattan distance between two points in p dimensions.(绝对距离) 今 When m=2,“ x, becomes the Euclidean distance.(欧氏距离) 2021/1/21
2021/1/21 23 cxt (2) Minkowski Metric(明考夫斯基距离) ❖ When m = 1, the measure called the “city-block” or Manhattan distance between two points in p dimensions.(绝对距离) ❖ When m = 2, becomes the Euclidean distance.(欧氏距离)
令明氏距离、欧氏距离以及绝对距离主要有以下两个 缺点: ①距离的值与各指标的量纲有关。各指标计量单位的 选择有一定的人为性和随意性,任何一个变量计量 单位的改变都会使此距离的数值改变,从而使该距 离的数值依赖于各变量计量单位的选择。 ②距离的定义没有考虑各个变量之间的相关性和重要 性。他们把各个变量都同等看待,将两个样品在各 个变量上的离差简单地进行了综合。 2021/1/21
2021/1/21 24 cxt ❖ 明氏距离、欧氏距离以及绝对距离主要有以下两个 缺点: ①距离的值与各指标的量纲有关。各指标计量单位的 选择有一定的人为性和随意性,任何一个变量计量 单位的改变都会使此距离的数值改变,从而使该距 离的数值依赖于各变量计量单位的选择。 ②距离的定义没有考虑各个变量之间的相关性和重要 性。他们把各个变量都同等看待,将两个样品在各 个变量上的离差简单地进行了综合
◆例:横轴代表重量(单位:kg,纵轴代表长度 (单位:cm)。有四个点ABCD见图。 10 AB=3+10=25 CD=√0+2=0l1 B 10 2021/1/21 25
2021/1/21 25 cxt ❖ 例:横轴 代表重量(单位:kg),纵轴 代表长度 (单位:cm)。有四个点A,B,C,D,见图。 C D A B 105 5 1 10 1 x 2 x 2 x 1 x 2 2 AB = + = 5 10 125 2 2 CD = + = 10 1 101