Assessing Normalization Techniques for Simple Additive Weighting Method

• Published in: Procedia computer science 2022-01, Vol.199, p.1229-1236
• Main Authors: Vafaei, Nazanin, Ribeiro, Rita A., Camarinha-Matos, Luis M.
• Format: Article
• Language: English
• Subjects: aggregation; Collaborative network; data fusion; Decision making; MCDM; Normalization; SAW
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2022.01.156

One of the current topics of attention in data analysis is the selection of the best normalization technique in the aggregation process when using Multi-Criteria Decision Making (MCDM) methods for solving decision problems. This is particularly critical in complex collaborative decision-making systems dealing with a large variety of heterogeneous data sources. Using different normalization techniques may result in different rankings of alternatives. So, enhancing the accuracy of the final ranking of alternatives could be achieved by selecting the most proper normalization techniques for each MCDM decision problem. In this direction, several attempts have been carried out, however, the lack of coherence and lack of a robust assessment framework persist. This situation encouraged the authors to propose an assessment framework that is enriched with several metrics for the evaluation of different normalization techniques in MCDM problems with the focus on partner/supplier selection in collaborative networks. As an illustration of the approach, in this work we assess different normalization techniques with the Simple Additive Weighting (SAW) method using metrics from the proposed assessment framework and select the most adequate technique for a small case study that is borrowed from literature. The suggested approach contributes to increasing the accuracy of final results for MCDM methods.