Cloud Service Recommendation System Using Knowledge Graph and PageRank Algorithm

The rapid growth of cloud computing technologies has resulted in a large number of cloud services and platforms across domains such as storage, networking, analytics, machine learning, and virtualization. Selecting suitable cloud services from a vast pool of available options has become a significant challenge for users and organizations. Traditional recommendation systems often fail to capture semantic relationships among cloud services and suffer from data sparsity, cold-start issues, and poor explainability. This paper proposes a lightweight and explainable Cloud Service Recommendation System using Knowledge Graph and PageRank Algorithm. The proposed system models cloud services and their semantic relationships using a Knowledge Graph and ranks related services using the PageRank algorithm. A real-world cloud services dataset containing 1,000 records, 467 unique products, 6 service categories, and 30 unique functionalities is used for evaluation. Unlike computationally intensive deep learning-based recommendation systems, the proposed approach provides efficient semantic ranking with lower computational complexity while maintaining recommendation quality and explainability. The system is implemented as an interactive web application using Streamlit and achieves a top PageRank score of 0.0038 for the highest-ranked cloud service. Experimental results demonstrate that the proposed system provides effective semantic recommendation with significantly lower computational overhead compared to Graph Neural Network-based approaches.