A PROPOSED METHOD FOR PREDICTING INCOMPLETE DESIGN ANALYTIC HIERARCHY PROCESS OBSERVATIONS IN A GROUP DECISION PROCESS

摘要

The proposed method estimates missing comparisons in an incomplete AHP design by regressing and predicting against one other individual in the group whose included scores correlate highly with those in question. Of particular interest is not only when two individuals have high positive correlation, but when they exhibit high negative correlation. The type of regression used is different for each. By estimating missing comparisons using other, similarly inclined individuals in the group, this method takes into consideration indistinct, subjective variables that may be specific to the uniqueness of the human respondent.