How to create a probability tree and calc prob of each branch?
6 views (last 30 days)
Show older comments
jeff bodington
on 1 Jul 2014
Commented: Roger Stafford
on 8 Jul 2014
I have six events and each event has a known probability of occurring first. All six occur but the order is unknown. There is thus one starting point and 6! (thus 720) ending points. I need to know the sequence of events and probability associated with each end point. This is probability tree 101. How do I set up such a tree in MATLAB (I do have Stat Toolbox)?
Thank you.
0 Comments
Accepted Answer
jeff bodington
on 2 Jul 2014
1 Comment
Image Analyst
on 5 Jul 2014
You can "thank" him by "Accepting his answer to give him reputation points which give him additional powers to do things in this forum.
More Answers (1)
Roger Stafford
on 2 Jul 2014
I assume that after the first event the following events occur with the same probabilities as the first events but conditioned on previous events not occurring. If so, do the following.
Let p be a vector of the probabilities of each of the events occurring first. These must sum to 1.
n = length(p);
r = perms(1:n); % Get list of all possible permutations
q = prod(p)./prod(cumsum(reshape(p(fliplr(r)),[],n),2),2);
Then q will be a list of the probabilities of each of the corresponding permutations in r - that is, the probability of each "end point". The values in q will sum to 1 (except for round-off error.)
2 Comments
Roger Stafford
on 8 Jul 2014
I suggest you break up the code for 'q' into the following steps and see at which step your trouble occurs. (I don't seem to have any trouble with it.)
t1 = fliplr(r);
t2 = p(t1);
t3 = reshape(t2,[],n);
t4 = cumsum(t3,2);
t5 = prod(t4,2);
t6 = prod(p);
t7 = t6./t5;
If n = 6, t1 should be a 720 x 6 array, t2 a 4320 x 1 vector, t3 a 720 x 6 array, t4 also 720 x 6, t5 a 720 x 1 vector, t6 a scalar (1 x 1), and t7 a 720 x 1 vector.
See Also
Categories
Find more on Image Segmentation and Analysis in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!