I am currently having a part of code using euclidean distance to compute some likelihoods. I am trying to think of any ways to compute likelihoods with the same effectiveness, without using euclidean distance. Is there anyone who could give me maybe a hint for a different way?
Asked
Active
Viewed 340 times
0
-
I don't know what exactly you are looking for; you could leave out the square root which is usually involved computing the distance. – Codor Nov 24 '15 at 13:59
-
Your question lacks sufficient context. In some contexts whatever formula you are using might be almost forced on you (e.g. if it is a maximum likelihood estimator for a particular distribution). In other contexts you are doing something more heuristic and have the freedom to experiment with other metrics such as the computationally easier taxicab metric. – John Coleman Nov 24 '15 at 14:13
-
In my first post I said that I want to find a different way to compute likelihoods with the same effectiveness with the euclidean, because as I found while searching, other methods like taxicab/manhattan metric doesn't always compute the min distance, while euclidean does. – userG Nov 24 '15 at 14:16