Merging the annotations from different annotators. The input
is the array containing the annotation set for merging. The
output is a map, the key of which is the merged annotations
and the values of which represent those annotators who agree
on the merged annotation. Two merging methods are implemented.
One method selects one annotation if at least a pre-defined
number of annotators agree on it. If there are more than
one merged annotations with the same span, the program selects only
one annotation from them with the maximal number of annotators
on it. Another method selects only one
annotation from those annotations with the same span,
which majority of the annotators support.