An Approach to Characterize Graded Entailment of Arguments through a Label-based Framework