The big O of of find-set depends on the implementation used.
As an example, take an array of your the data in the set. They are sorted, your can access them them with O(log(N)) (or even O(log(log(N)) but then any insertion is O(N) -- think about you add a new lowest element into our set. You can improve this by using heaps.
Or you store the set in a linked list. Adding and deleting is O(1), but to find it, you'll have to traverse all entries, thus O(N).
You have to decide, which operation is used more often or in a more crucial point of your algorithm.