normalized conditional mutual information

Asked Jul 07 '21 at 12:55

Active Jul 07 '21 at 13:02

Viewed 244 times

I am learning mutual information and need to apply it to my data. I strongly need to find a normalized conditional mutual information to have a range within [0,1] instead of 0 to infinity in conditional mutual information. I search about such a problem and could not find any idea how to find the solution using, for example, the R program. I just wonder if there is any way to explain what is the normalized conditional mutual information and how to get it?

edited Jul 07 '21 at 13:02

Arya McCarthy

8,787

asked Jul 07 '21 at 12:55

Maryam

1,620

2

Why do you need a normalized value? // You might be interested in a question of mine from last year. – Dave Jul 07 '21 at 13:13
@Dave Thanks a lot for your comment. I need it to have a finite value of the mutual information instead of getting the range from 0 to infinity. I have read your question and answers but I still cannot find any R package that can deal with conditional mutual information or even conditional adjusted mutual information as I really need to calculate the conditional of two variables given other variables. – Maryam Jul 07 '21 at 15:55
The value of mutual information you get when you do the calculation will be finite. – Dave Jul 07 '21 at 15:56
do you mean the adjusted mutual information? how about conditional value? – Maryam Jul 07 '21 at 15:57
Could you please show me an example. – Maryam Jul 07 '21 at 16:00
You are adding up a finite number of finite quantities, so the result will be finite. – Dave Jul 07 '21 at 16:01
@Dave Thank you a lot for your help and time. I do have conditional values of more than 1 and I need them to be between 0 and 1. That will be easy for me to interpret as a strong or weak association. – Maryam Jul 07 '21 at 16:08
Let us continue this discussion in chat. – Dave Jul 07 '21 at 16:15
@Dave Yes sure. I am in the chat. – Maryam Jul 07 '21 at 16:36

normalized conditional mutual information

0 Answers0