Suppose you have attributes 10 attributes in a relation R as below

R(A, B, C, D, E, F, G, H, I, J )

Now you start finding the candidate key.

Candidate key:- The mininal set of attribute which can uniquely determine all the other attributes.(Note - NOT Minimum set).

You started by taking one attribute.

Suppose B can determine all the other (9) attributes then B is our candidate fine. (It's minimal)

Now, you know that any superset of B will also determine all the other attributes[ Consider AB, It will determine all the attributes because of B.

Now it can not be the candidate key because it's minimal version that it B, is already a candidate key ]. So now we can check taking 2 attributes at a time [example, CD , AD, IJ,....](but not BD, BA...).

You found that { GH, IJ } both the pair can determine all the other attributes. So both GH and IJ will also be our candidate keys. {Note- these are minimal version of itself because you can not remove I or H}.

And so on.