I would like to commend the authors for their insightful work entitled “Intelligent Decision Support Systems—An Analysis of Machine Learning and Multicriteria Decision-Making Methods.” This paper tackles a pertinent and increasingly important topic, offering a detailed exploration of how intelligent systems can leverage machine learning (ML) and multicriteria decision-making (MCDM) methods to address complex decision-making challenges across various domains. However, I have several observations and questions regarding the study. My remarks aim to foster constructive discussion and further refinement of this work in a promising research area.
The manuscript provides a comprehensive review of MCDM methods and their integration with ML for intelligent decision support systems (IDSS). While the theoretical and conceptual frameworks are well articulated, there are notable concerns regarding the methodological transparency and empirical grounding of the study. The selection of literature, although broad, does not sufficiently justify its inclusion criteria, and key details about the quality assessment process remain ambiguous. This lack of transparency limits the reproducibility and reliability of the findings.
Additionally, while the authors highlight numerous application areas for IDSS, such as agriculture, medicine, and transportation, the analysis occasionally veers into overgeneralization. Claims about the effectiveness of these systems in addressing domain-specific challenges are not consistently supported by empirical evidence or detailed case studies. For instance, the study cites the benefits of ML-enhanced decision-making in medicine and agriculture but fails to provide concrete examples or real-world implementations that validate these claims. Including such examples would greatly enhance the practical relevance and credibility of the findings.
The review methodology, particularly the systematic literature review (SLR), is described but lacks critical detail. The study mentions the use of logical operators in query formulation and filtering but does not explain how the relevance of retrieved papers was determined beyond titles and abstracts. Furthermore, the exclusion of certain libraries and potential language bias (e.g., focusing solely on English publications) raises questions about the comprehensiveness and objectivity of the review.
The discussion of MCDM methods and their integration with ML is insightful but could benefit from a more nuanced analysis of limitations and trade-offs. For example, the scalability of these methods, their computational requirements, and their performance under real-world constraints (e.g., noisy or incomplete data) are not sufficiently explored. A critical evaluation of these factors is essential to provide a balanced perspective and avoid overestimating the feasibility of the proposed approaches.
Moreover, the study’s claims about the impact of IDSS across various domains are compelling but occasionally unsupported by robust evidence. Quantitative assertions, such as improvements in decision-making efficiency or effectiveness, are often presented without accompanying metrics, baselines, or statistical validation. This undermines the strength of the conclusions and highlights the need for more rigorous empirical validation.
The cited literature, while extensive, exhibits some noticeable gaps. Influential works on ML and MCDM applications, particularly those addressing emerging challenges and novel methodologies, are conspicuously absent. Addressing these gaps would provide a more comprehensive understanding of the field and better situate the study within the broader academic discourse.
Lastly, the paper would benefit from improved organization and clarity. Certain sections are repetitive or overly verbose, which detracts from the manuscript’s readability and impact. A more concise presentation, emphasizing actionable insights and practical implications, would enhance the accessibility and utility of the paper to both researchers and practitioners.
I look forward to the authors’ responses to these points and their thoughts on addressing these limitations. I believe that refining these aspects will significantly strengthen the paper’s contribution to the field, advancing our understanding of intelligent decision support systems and their integration with machine learning and multicriteria decision-making methods.