Basic Statistical Methods in Determining Criteria Weights

Abstract

The proliferation of technology has facilitated data accessibility, leading to an expansion in the range of criteria employed in decision problem design. This situation offers an advantage for making precise and rational decisions, but when it comes to managing spending, it becomes a disadvantage. Specifically, the expense of acquiring expert views utilized in the computation of criteria weights by subjective approaches experiences a substantial rise. Hence, decision-makers may employ objective methodologies to determine criterion weights. Nevertheless, objective methods provide a more limited range of choices compared to subjective methods. The study aims to utilize two widely recognized fundamental statistical approaches in order to enhance the capabilities of objective methods. One of the suggested approaches is the dissimilarity-based weighting method, which calculates the differentiation of values within the criteria. Another approach is the weighting method, which relies on the interquartile range. The methods were adapted as means of weighting criteria. Explanatory examples were provided, simulation-based comparisons were conducted, and ultimately applied to an actual data set. The data from each scenario were compared using the factorial analysis of variance method. The findings produced demonstrate that the proposed methods align with other objective methodologies. Furthermore, the proposed approaches were observed to take more time to finish the procedure compared to the Entropy and Standard Deviation methods, but less time compared to the Critic and Merec methods. Consequently, the suggested techniques are introduced as alternative approaches derived from established fundamental statistical procedures, which are straightforward to comprehend and valuable for professionals.