find probability of chi2 test

Dear all;

if i want to perfrom a hypothesis test for chi2 distrubtion , how can i do it in matlab :

P(X>a)= alpha

[Asvin Kumar]

If you're looking for the chi-squared test, have a look at chi2gof.

[Mustafa Al-Nasser]

Thank you

i need to check P(x>Chi2_critical Value)=alpha
 alhpha can be specifed but how i can specfiy Chi2_critical Value , usually i am looking at degree of freedom of 2 
 so, Chi2_critical Value = 5.99 at alpha= 0.05

[Asvin Kumar]

I'm not sure I can help much with this topic. Just curious... If Chi2_critical and alpha are constants, what is x? What distribution does it follow? Do you have a sample set of values of x?

If there's a wiki page explaining this statistic, could you please link it? I'd like to learn something new.

[Mustafa Al-Nasser]

thank you Asvin

basically i need to test the following hypothesis

P(X> Chi2_Cirtical value)= alpha

Chi2_Cirtical value calcuated from central Chi2 distribution table , alpha is the signifacnce level

so, i need to check if the hypothesis hold if the above equatio is satisifed .

let us assume that Chi2_Cirtical value= 5.99 at alpha= 0.05 .at degree of freedom 2

in chi2gof. , i can specify the alpha but to specify the Chi2_Cirtical value , I need to know the degree of freedom.

if i use above function , i can got degree of freedeom , so since i know alpha i, i can check if the dgree of freedom > 2 , then X> 5.99 and my ypothesis is hold.

Is my conclusion is right?

[Asvin Kumar]

Thanks for explaining. I'm, unfortunately, not the right person for this question. I was merely suggesting a function I had come across in the docs. I'll let someone else from the community help you with this.

Answer 1

[Ive J]

Based on what you explained it seems you already have the X2 stat.

p = 1 - chi2cdf(x, nu) % where x is chi2 stat and nu is df

Answer 2

[Aditya Patil]

To elaborate on @Ive J's answer,

You have X. using chi2cdf, you can get . So in your case, , which you can get using chi2cdf as follows

p = 1 - chi2cdf(X, a)