CDF for Gaussian Mixture Distribution

Question

dsmalenb 2016-3-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/274578-cdf-for-gaussian-mixture-distribution

编辑： dsmalenb 2016-3-20

Hello,

I am ultimately trying to calculate the KS-test statistic for a mixture Gaussian distribution for a data set 'TestData' I have tried to find the appropriate syntax for this but I cannot seem to get it correct. Below is my code with the error messages. I have also included my 1 mixture Gaussian solution as a guide. I didn't think it would be much different. I still think it might not be. Thank you for any help you cn provide.

% Mixture Code:

for j=1:10000 if rand <=0.4 TestData(j) = normrnd(0,2); else TestData(j) = normrnd(10,2); end end

% Fit TestData to Gaussian 2 Mixture Model mpd01 = fitgmdist(TestData, 2);

% Calculate the CDF value for the Mixture Distribution at a given TestData % value test_cdf = [TestData,cdf(TestData,mpd01.mu, mpd01.sigma, mpd01.p)];

% Calculate hypothesis test and p-value for the KS test [h,p] = kstest(test_cdf, 'Alpha', 0.01);

% --------------- Error Message

%No appropriate method, property, or field 'sigma' for class 'gmdistribution'.

%Error in classreg.learning.internal.DisallowVectorOps/subsref (line 21) % [varargout{1:nargout}] = builtin('subsref',this,s);

%Error in BurrTest (line 10) %test_cdf = [TestData,cdf(TestData,mpd01.mu, mpd01.sigma, mpd01.p)]

% Singular Gaussian Code

% Normal Distribution

pd18 = fitdist(TestData,'normal'); test_cdf = [TestData,cdf('normal',TestData,pd18.mu,pd18.sigma)]; [h,p] = kstest(test_cdf, 'Alpha', 0.01); pValKStest(18) = p;