Hi Jakir,

The following function in the File Exchange does 2-d kernel density estimation:

http://www.mathworks.com/matlabcentral/fileexchange/17204-kernel-density-estimation

It is highly rated but I have not tried it yet. Hope, this is helpful.

Jakir,

I would suggest determining whether your variables are discrete/continuous and dependent/independent first, then proceed to calculate their joint pdf. Regardless, the wikipedia page provides the definition of the joint pdf of n discrete random variables as:

f(X1,X2,...,Xn)= f(X1=x1)*f(X2=x2|X1=x1)*f(X3=x3|X1=x1,X2=x2)* .... * f(Xn=xn|X1=x1,X2=x2,...,Xn-1=xn-1)

https://en.wikipedia.org/wiki/Joint_probability_distribution

Hope this helps.

Thank you @Ganesh Naik, I already have tried this method, I can also calculate Joint PDF upto 3 variables using mvnpdf() function in MATLAB. But i need joint pdf for more than 3 variables.

Regards,

Jakir

Thank you @Arman Niaki, I know the theories. the problem is implementing them in MATLAB. There is no direct way to calculate those conditional probability from data.

Regards,

Jakir

Jakir,

Theoretically speaking, you should be able to edit the script or re-write the function: mvnpdf() that would evaluate each conditional probability (i.e. what I mentioned before: f(X1,X2,...,Xn)= f(X1=x1)*f(X2=x2|X1=x1)*f(X3=x3|X1=x1,X2=x2)* .... * f(Xn=xn|X1=x1,X2=x2,...,Xn-1=xn-1)

then, f(X2=x2|X1=x1)= f(X2=x2 intersect X1=x1)/f(X1=x1) )

However, this is most definitely going to be computationally heavy for Matlab to handle, especially if you have alot of RV's.

I would suggest to look into using Mathematica as it has out performed Matlab and Python computationally in my personal experience.

Hope this helps,

Hello Look for the probability density function in help, it depends of course on the distribution Good luck