Ndtyjky

I thought there would be a library that would help me to do this task instead of writing many lines of codes. I tried finding some solutions from books related to my problem, but I could not find any.

One of the recent books I did read related to probability:

Python for Probability, Statistics, and Machine Learning for José
Unpingco

The task is that I have a matrix like this one below

0    1

213 vha

342 gha

523 xha

121 gha

812 gha

612 vha

123 gha

and I want the program to calculate the steps of moving from, say, vha to gha in the second row. and from gha to xha in the third row. If any step is repeated, it will be added to the previous step. For example vha to gha in the first and second rows is repeated at the end of the matrix.

The desired output is will be the similar steps added together/ total number of rows-1. In the first case it is vha to gha prob = 2/7-1

Desired output

vha to gha prob = 0.3

gha to xha prob = 0.16

xha to gha prob = 0.16

gha to gha prob = 0.16

gha to vha prob = 0.16



Total probs = 1

You can use a Counter to count how many times a transition occurs and then calculate probabilities for each transition.

You can use zip to combine two slices of the list m - one with the last element removed and another with the first element removed - to get tuples for adjacent elements. zip(m[:-1], m[1:]) does that. Then can you can count similar tuples - which represent transitions - with a Counter:

from collections import Counter



m = [[213, 'vha'],

     [342, 'gha'],

     [523, 'xha'],

     [121, 'gha'],

     [812, 'gha'],

     [612, 'vha'],

     [123, 'gha']]



c = Counter([(x[1], y[1]) for x, y in zip(m[:-1], m[1:])])

probs = [(e, v / (len(m) - 1)) for e, v in c.items()]



for p in probs:

    print(p)

Output

(('vha', 'gha'), 0.3333333333333333)                                                                                                                

(('gha', 'xha'), 0.16666666666666666)                                                                                                               

(('xha', 'gha'), 0.16666666666666666)                                                                                                               

(('gha', 'gha'), 0.16666666666666666)                                                                                                               

(('gha', 'vha'), 0.16666666666666666)