Krishnan Anath Rau
Labuan Faculty of International Finance, Universiti Malaysia Sabah, Labuan, Malaysia.
Front Big Data. 2022 Aug 18;5:990699. doi: 10.3389/fdata.2022.990699. eCollection 2022.
The use of a multi-criteria decision-making (MCDM) technique mostly begins with normalizing the incommensurable data values in the decision matrix. Numerous normalization methods are available in the literature and applying different normalization methods to an MCDM technique is proven to deliver varying results. As such, selecting suitable normalization methods for an MCDM technique has emerged as an intriguing research topic, especially with the advent of big data. Several efforts have been made to compare the suitability of various normalization methods, but regrettably, no paper provides an updated review of these crucial efforts. This study, therefore, aimed to trace articles reporting such efforts and review them based on the following three perspectives: (1) the normalization methods considered, (2) the MCDM methods considered, and (3) the comparison metrics used to determine the suitable normalization methods. The relevant articles were extracted with the aid of Google Scholar using the keywords of "normalization" and "MCDM," and Tableau software was used to analyze further the data gathered through the articles. A total of five limitations were uncovered based on the current state of literature, and potential future works to address those limitations were offered. This paper is the first to compile and review the previous investigations that compared and determined the ideal normalization methods for an MCDM technique.
多准则决策（MCDM）技术的使用通常始于对决策矩阵中不可通约的数据值进行归一化处理。文献中存在多种归一化方法，事实证明，将不同的归一化方法应用于MCDM技术会产生不同的结果。因此，为MCDM技术选择合适的归一化方法已成为一个引人关注的研究课题，尤其是在大数据出现之后。人们已经做出了一些努力来比较各种归一化方法的适用性，但遗憾的是，没有论文对这些重要的努力进行更新的综述。因此，本研究旨在追踪报道此类努力的文章，并基于以下三个视角对其进行综述：（1）所考虑的归一化方法；（2）所考虑的MCDM方法；（3）用于确定合适归一化方法的比较指标。借助谷歌学术，使用“归一化”和“MCDM”作为关键词提取相关文章，并使用Tableau软件进一步分析通过这些文章收集的数据。基于当前的文献状况共发现了五个局限性，并提出了应对这些局限性的潜在未来工作。本文首次汇编并综述了之前比较和确定MCDM技术理想归一化方法的研究。
