i want to compute a cdf for t distribution

1 次查看(过去 30 天)
i want to compute a cdf for t distribution from a to b , i wrote this code but it didn't run
for i=1:50
fa(i)=cdf('t', 'mu' =-2, 'sigma'= 4 , 'nu'= 6 , astar(i))
fb(i)=cdf('t', 'mu' =-2, 'sigma'= 4 , 'nu'= 6, bstar(i))
end
astar and bstar are points of the integral
sorry for a weak my language

采纳的回答

Star Strider
Star Strider 2018-3-1
You are not calling cdf correctly. Also the functions are vectorised, so you do not need the loops.
Try this:
fa = cdf('T', nu, astar);
fb = cdf('T', nu, bstar);
It is probably best to use the tcdf (link) function instead, since cdf appears to be deprecated in a future release of MATLAB.
  2 个评论
elham kreem
elham kreem 2018-3-1
thanks for you , your answer for stander dist. and i apply it as : fa = cdf('T', 'nu', 6, astar) the result is : fa =
1.0000 1.0000
but if i want cdf for non stander distribution for t, I apply this code
fa=cdf('T', 'mu' ,-2, 'sigma', 4 , 'nu', 6 , astar)
but the result is :
fa =
NaN NaN
Star Strider
Star Strider 2018-3-1
My pleasure.
In the t-distribution, ‘nu’ is calculated as described in Compute Student's t cdf (link). You use the mean (‘mu’), standard deviation (‘sigma’) and number of observations (‘n’) to calculate it. There is only one parameter.
If you use my code, you get the correct result:
astar = linspace(1E-5, 2, 5);
nu = 6;
fa = cdf('T', nu, astar)
I again urge you to use the tcdf function (that I linked to in my Answer), instead of cdf.

请先登录,再进行评论。

更多回答(0 个)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by