Scholarlink

<p>Existing Cloud service selection models have often focused solely on either direct (performance-based) or indirect (feedback-based) reputation assessment, leading to incomplete and potentially unreliable service recommendations. The challenge is to integrate these diverse data streams into a single, robust framework for comprehensive reputation management and trustworthy service selection. The framework’s design utilizes a linear programming algorithm in Module 1 for efficient overall ranking, employs the QR matrix decomposition algorithm in Module 2 to quantify direct reputation from real-time Cloud services monitoring (analysing parameters like availability, reliability, etc), and uses a Naïve Bayes technique enhanced by an LSTM model in Module 3 to handle indirect reputation derived from web-scraped customer feedback and service cost. The major finding is that this comprehensive system provides a powerful, interpretable, and superior ranking approach that effectively calculates reputation and trust values, which are systematically validated against existing models. In conclusion, this unique combination of assessment methods offers a holistic solution to Cloud service selection, despite acknowledged limitations regarding the Naïve Bayes assumption and data acquisition, leading to future work focused on advanced AI models, dynamic weighting, and blockchain integration for enhanced transparency and accuracy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A multi-module framework for Cloud service reputation management and selection

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • R. Devi

摘要

Existing Cloud service selection models have often focused solely on either direct (performance-based) or indirect (feedback-based) reputation assessment, leading to incomplete and potentially unreliable service recommendations. The challenge is to integrate these diverse data streams into a single, robust framework for comprehensive reputation management and trustworthy service selection. The framework’s design utilizes a linear programming algorithm in Module 1 for efficient overall ranking, employs the QR matrix decomposition algorithm in Module 2 to quantify direct reputation from real-time Cloud services monitoring (analysing parameters like availability, reliability, etc), and uses a Naïve Bayes technique enhanced by an LSTM model in Module 3 to handle indirect reputation derived from web-scraped customer feedback and service cost. The major finding is that this comprehensive system provides a powerful, interpretable, and superior ranking approach that effectively calculates reputation and trust values, which are systematically validated against existing models. In conclusion, this unique combination of assessment methods offers a holistic solution to Cloud service selection, despite acknowledged limitations regarding the Naïve Bayes assumption and data acquisition, leading to future work focused on advanced AI models, dynamic weighting, and blockchain integration for enhanced transparency and accuracy.