Comments on "A new combination of evidence based on compromise" by K. Yamada

Basically, most combination rules are mainly based on a conjunctive operator and propose a specific way of redistributing the global (or the partial) conflicting belief mass among some elements of the power-set (or the hyper-power set) of the frame of discernment. Using the conjunctive rule means that if the experts agree (i.e. their testimonies have a non-empty intersection), we consider them as reliable and if they are in conflict (empty intersection), at least one of the experts is considered unreliable see Dubois and Prade (1988). Then, the disjunctive combination rule can be employed instead see Dubois and Prade (1986). However the disjunctive rule is generally not used because it deteriorates the specificity of the expert’s responses, i.e. the combined mass is usually less specific after the disjunctive fusion than the mass of each source taken separately. If the reliability of the experts is unknown, there are different techniques to estimate it Elouedi et al. (2004); Martin et al. (2008).