Enhanced cloud service ranking using a hybrid RoBERTa-based deep learning framework with multi-criteria decision support

CSPs vary widely in terms of performance, security, and user satisfaction and one must have the right ranking. This research paper hypothesizes a Hybrid RoBERTa-based deep learning model combined with Multi-Criteria Decision Analysis (MCDA) to conduct a strong evaluation of cloud services. This framework is a summary of structured metrics (latency, availability, throughput, security) and unstructured data (user reviews and incident reports) of major CSPs, such as AWS, Microsoft Azure, and GCP. The Hybrid RoBERTa model shows better results compared with traditional methods (SVM, Naive Bayes, LSTM, and BERT) and its simulation shows that the approach has an accuracy of 97.25, preci-sion of 0.976, recall of 0.976, and F1-score of 0.972. The user sentiment, latency, and availability were found to be the most effective factors using the feature importance analysis. With the TOPSIS method, AWS was ranked the first position with Mi-crosoft Azure coming second and Google Cloud at the third position. The findings provide support of the model proposed by the authors in providing precise, dynamics, and security-conscious rankings of cloud services, which can serve as a scalable basis of future AI-oriented decision support systems.