原文:http://blog.sina.com.cn/s/blog_7054a1960102vy7x.html
累积分布函数cdf (Cumulative Distribution Function)
背景知识:http://www.lifelaf.com/blog/?p=746
语法
y = cdf('name',x,A,B)
y = cdf('name',x,A,B,C)
y = cdf(pd,x)
y = cdf(___,'upper')
描述
y = cdf('name',x,A) 计算某种分布(由'name'定义,如'Normal'正态, 'Poisson'泊松, 'T' t分布…)下,x值处的累计分布,A,B,C等为'name'函数的参数
y = cdf(pd,x) 直接计算概率分布函数pd(probability distribution) ,在x处的累计分布,实际上,这里的pd 已被'name', A定义好,举栗如下:
% 定义一个正态分布函数pd, 均值mu = 0, 标准差sigma = 1.
mu = 0;
sigma = 1;
pd = makedist('Normal',mu,sigma);
% 定义x值
x = [-2,-1,0,1,2];
% 计算x值处的累计分布
y = cdf(pd,x)
y =
0.0228 0.1587 0.5000 0.8413 0.9772
用第一种语句表达相同内容为:
y2 = cdf('Normal',x,mu,sigma) %正态分布,x值处,均值为0,标准差为1
y2 =
0.0228 0.1587 0.5000 0.8413 0.9772
http://nl.mathworks.com/help/stats/cdf.html
t分布累积分布函数tcdf (Student'stcumulative distribution function)
% 事实上就是y = cdf('T',x,A)函数
语法
p = tcdf(x,nu)
p = tcdf(x,nu,'upper')
描述
计算t分布在x值处的累积分布,nu是t分布的自由度
再举个栗子
mu = 1; % Population mean
sigma = 2;% Population standard deviation
n = 100; % Sample size
x = normrnd(mu,sigma,n,1);% Random sample from population
xbar = mean(x);% Sample mean
s = std(x); % Sample standard deviation
t = (xbar - mu)/(s/sqrt(n)) % 这里t分布出现了,正态分布总体与样本均值的差符合t分布
t =
1.0589
p = 1-tcdf(t,n-1) % Probability of larger t-statistic
p =
0.1461
该p值(即t函数的累积分布就是t检验在相同x值处的概率ptest)
[h,ptest] = ttest(x,mu,0.05,'right')
h =
0
ptest =
0.1461
http://nl.mathworks.com/help/stats/tcdf.html
概率密度函数pdf (Probability density functions)
搞懂了累积分布函数cdf,这个就没什么需要多说了
语法
y = pdf('name',x,A)
y = pdf('name',x,A,B)
y = pdf('name',x,A,B,C)
y = pdf(pd,x)
举例
% 定义一个正态分布函数pd, 均值mu = 0, 标准差sigma = 1.
mu = 0;
sigma = 1;
pd = makedist('Normal',mu,sigma);
% 定义x值
x = [-2 -1 0 1 2];
% 计算x值处的概率密度(cdf是累计分布)
y = pdf (pd,x)
y =
0.0540 0.2420 0.3989 0.2420 0.0540
同样,另一种表达
y = pdf(pd,x)
y =
0.0540 0.2420 0.3989 0.2420 0.0540
http://nl.mathworks.com/help/stats/pdf.html
t分布概率密度函数tpdf(Student's t probability density function)
语法
y = tpdf(x,nu)
举例
tpdf(0,1:6)
ans =
0.3183 0.3536 0.3676 0.3750 0.3796 0.3827
http://nl.mathworks.com/help/stats/tpdf.html
相反,还可以通过p求t分布的t值
tinv (Student's t inverse cumulative distribution function)
语法
x = tinv(p,nu)
举例
% the 99th percentile of the Student's t distribution for one to six degrees of freedom
percentile = tinv(0.99,1:6)
percentile =
31.8205 6.9646 4.5407 3.7469 3.3649 3.1427
http://nl.mathworks.com/help/stats/tinv.html
有一个问题,Matlab有一个inv矩阵求逆函数,不知与tinv什么关系,莫非tinv是在t分布下调用了inv计算程序?但p并不等是t的逆矩阵啊(即t*p = E)啊?求解答
inv是矩阵求逆的意思。具体用法A=inv(B),其中B是输入的可逆矩阵,输出A就是B的逆矩阵,逆矩阵满足性质 AB=BA=E (E是单位阵)。如果输入的是不可逆矩阵会弹出警告,并返回inf。
调用举例:
>> inv([1 0;0 0])
警告: 矩阵为奇异工作精度。
ans =
Inf Inf
Inf Inf
>> inv(rand(2))
ans =
-13.0929 5.2640
12.0501 -3.3159
另附官方英文解释(输入doc inv也可以自己查看):
Y = inv(X) returns theinverse of the square matrix X. A warning messageis printed if X is badly scaled or nearly singular.
In practice, it is seldom necessary to form the explicit inverseof a matrix. A frequent misuse of inv arises whensolving the system of linear equations Ax = b.One way to solve this is with x = inv(A)*b.A better way, from both an execution time and numerical accuracy standpoint,is to use the matrix division operator x = A\b.This produces the solution using Gaussian elimination, without formingthe inverse. See mldivide (\)for further information.
网友评论