Not to repeat the same thing, please just read the question in: https://mathematica.stackexchange.com/questions/179613/constructing-transition-probability-matrix and the answer by @Anton Antonov and @kglr. I like to create a transition matrix using a weighted, directed graph with a domain of weights in the interval [0, 1.5]. My states will be randomly chosen intervals. I may divide the domain into equal sub-intervals or different-length intervals at my choice. The transition matrix will consist of the number of binary links in each subset. The following weighted graphs with 5 vertices can be used to answer. SeedRandom[0]; matT1=RandomReal[{0,1.5}, {5,5}]; (* time t *) matT2=RandomReal[{0,1.5}, {5,5}]; (* time t+1 *) **EDIT 1** 1. Suppose that I have three states at time t: `{s1, s2, s3}` represented respectively by three intervals of edge weights: `{r1, r2, r3}` where `r1=[0, 0.3]`, `r2=(0.3, 0.6]`, `r3=(0.6, 3]` with, say, max weight is 3 in both graphs. 2. Find the number and actual linkages in each interval at t and then by using time t+1 digraph, identify the number of linkages that start from r1 at t and moving to r2 and r3 at t+1 and remaining at r1; We need to repeat the same procedure for 3 states to find out the final transition matrix. 3. Eventually, we need to derive a transition matrix using two digraphs that reflect the transition of linkages from one interval at t to another at t+1 (a typical transition matrix). 4. Additionally, I like to get the list of binary linkages themselves to analyze the set of linkages associated with a limiting probability. If the limiting probability is 0.25 for `s1` then I like to know which linkages are more likely to be in that state. I hope my explanation is clear.