Mutual Information between continuous random variables

I want to determine the mutual information or information gain between the 30 years rainfall data at two stations. Since, rainfall data is a continuous variables, can someone suggests me a matlab code to determine the mutual information?

[KSSV]

What do you mean by mutual information?

[Deepthi B]

https://quantdare.com/what-is-mutual-information/

[Swarnalee Mazumder]

Hi @Deepthi B, I am trying to do something similar. I was wondering if this issue has been solved or not.

Answer 1

[KSSV]

https://in.mathworks.com/matlabcentral/fileexchange/28694-mutual-information

https://in.mathworks.com/matlabcentral/fileexchange/13289-fast-mutual-information-of-two-images-or-signals

https://in.mathworks.com/matlabcentral/fileexchange/35625-information-theory-toolbox

https://in.mathworks.com/matlabcentral/fileexchange/10040-average-mutual-information

[Deepthi B]

the problem is that these codes are all for discrete variables . I want it for continuous variables

[Jeff Miller]

Tough problem, because you need the density of the joint distribution to compute the mutual information for continuous variables. Unless you have a joint distribution family in mind (e.g., bivariate normal) and a way to estimate its density (easy for normal), then (as far as I can see) you will have to categorize your continuous scores into discrete bins and then use the discrete methods.

[Deepthi B]

Can you provide some materials related to this if you don't mind?

[Jeff Miller]

Sorry, I have no related materials to provide.

[Deepthi B]

Okay Sir. Thank You

[Deepthi B]

@Jeff Miller I don't know whether it is right to ask here. Can we do it in Python? Sorry if this question doesn't suits here

[Jeff Miller]

I don't know Python, but I doubt it. I don't think you can do anything without knowing or assuming the underlying joint distribution family, regardless of the programming language.