Integrating multicriteria decision making and principal component analysis: a systematic literature review

Abstract

Decision-support methods are crucial for analyzing complex alternatives and criteria in today’s data-driven world. This Systematic Literature Review (SLR) explores and synthesizes knowledge about decision support methodologies that integrate Multicriteria Decision Making (MCDM) and Principal Component Analysis (PCA), an unsupervised Machine Learning (ML) technique. Both techniques optimize complex decisions by combining multiple criteria and dimensional data analysis. Focusing on performance evaluations, criterion weighting, and validation testing, this review identifies significant gaps in existing methodologies. These include the lack of consideration for non-beneficial criteria in PCA, insufficient validation tests in over half of the studies, and the non-use of communalities (the contribution of each criterion to the main factors) in decision support approaches. Additionally, this SLR offers a comprehensive quantitative overview, analyzing data from the Scopus, IEEE, and Web of Science databases and identifying 16 relevant studies. Furthermore, the scarcity of systematic reviews integrating MCDM and PCA techniques impedes evidence-based decision-making practices and theoretical evolution. This is particularly crucial as ML and data analysis advance rapidly, requiring models that reflect technological innovations. This article addresses this gap in the literature by providing an analysis of decision support methods and guiding further improvement in this field.

Keywords

• Multicriteria decision-making
• principal component analysis
• decision support methods
• machine learning
• Dr Jenhui Chen, Computer Science and Information Engineering, Chang Gung University, Taoyuan, Taiwan
• Applied Mathematics