Hybrid Intelligence for Cloud Service Evaluation: Integrating PCA-Feature De-correlation and Best Only Weight Generation

Abstract

Choosing a cloud service has become a very important multi-criteria decision-making (MCDM) problem because there are more and more providers that offer different quality-of-service (QoS) features. Conventional MCDM methodologies, including the Analytic Hierarchy Process (AHP) and the Best–Worst Method (BWM), frequently encounter challenges related to inconsistency, subjectivity, and scalability, particularly when addressing numerous criteria and alternatives. This study introduces an innovative hybrid framework that combines Principal Component Analysis (PCA) with the Best-Only Method (BOM) to achieve efficient and consistent ranking of cloud service providers (CSPs). PCA is used to get rid of duplicate and correlated QoS criteria, which makes dimensionality reduction more objective. BOM, on the other hand, uses the fewest pairwise comparisons possible to find relative criteria weights, which makes the process more efficient. The integrated PCA-BOM method is also linked to a ranking system that lets you compare different CSP options. The experimental results show that the proposed framework has a better rank conformance with the baseline BOM, a perfect consistency ratio (CR = 0.00), and a high Ranking Stability Index (RSI), which means it is strong in different decision-making situations. It also needs up to 83% fewer comparisons than AHP and 48% fewer than BWM, which shows that it can be used in practice. The hybrid PCA-BOM framework is a structured, reliable, and computationally efficient way to rank cloud services based on more than one criterion.