User‐Centric and Expertise‐Driven Cloud Service Recommendation Approach

ABSTRACT Selecting the most appropriate cloud service remains a challenge due to the large number of available options and the diverse preferences of users. Traditional recommendation methods have primarily relied on Quality of Service (QoS) metrics, often neglecting the depth and reliability of user‐generated feedback. However, not all user reviews have the same value: Expert users provide more informed and relevant information, making expertise a key factor in improving the accuracy and trustworthiness of recommendations. This paper proposes an expertise‐driven recommendation approach that integrates Convolutional Neural Networks (CNN) for sentiment classification with an expertise‐weighted ranking mechanism. User expertise is assessed through multiple factors, including technical terminology usage, review credibility, and community validation, ensuring that recommendations are guided by reliable and well‐informed user contributions. By leveraging expertise‐Based recommendations, the approach enhances personalization and robustness, overcoming the limitations of traditional QoS‐based, trust‐based, and popularity‐based methods. Experimental results confirm the effectiveness of expertise‐aware recommendation, demonstrating its ability to address key limitations of existing recommendation approaches. This research contributes to the advancement of adaptive, user‐centric, and expertise‐aware cloud service recommendation models, promoting more reliable and context‐aware recommendations in dynamic cloud environments.