I have several positive integer numbers representing the count of features, for example delay, successful_attempts, failed_attempts. Say I have 2 sets of such numbers, A and B:
A : delay=63, successful_attempts=6, failed_attempts=7
B : delay=84, successful_attempts=5, failed_attempts=12
I am trying to find a metric to use in order to find how each feature counter between A and B differs. That is, a metric which will compare the delays of A and B and give a number representing their distance. Similarly for the other two feature counters.
And finally, deduce a total difference of how similar the counters between A and B are, ideally via a percentage and with the option -if possible- of adding weights to each feature.
I have searched for such metrics but could not find anything; probably cause I was not searching correctly, I was getting hits for string comparison mainly.
I have considered the Euclidean Distance and the Absolute distance but would appreciate if someone could confirm if these are the appropriate metrics to use for my case and/or let me know of any other metrics more suitable.