Scholarlink

<p>The Internet of Things (IoT) is experiencing rapid growth, with an increasing number of connected devices generating large volumes of data, which poses challenges for real-time processing and management. Edge computing provides a viable solution by enabling local data processing and reducing reliance on centralized data centers, thereby decreasing latency. However, data redundancy remains a key challenge that can negatively impact system performance. This study proposes an edge computing optimization framework that integrates K-Means clustering with decision tree-based data analytics. The proposed approach aims to reduce redundant data and improve data handling efficiency at the network edge. Simulation results indicate improvements in system performance compared to existing methods, including an increase in overall efficiency by 12%, an improvement in data analysis accuracy by 9%, a reduction in energy consumption by 11%, and a decrease in response time by 7%.</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 hybrid data-driven framework for real-time edge computing optimization in IoT environments

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月22日

查看全文
  • Qing Zhang,
  • Nahideh Derakhshanfard,
  • Ghazaleh Pourmahmood,
  • Amir Mollanejad,
  • Sana Zarrinpour,
  • Mehdi Hosseinzadeh

摘要

The Internet of Things (IoT) is experiencing rapid growth, with an increasing number of connected devices generating large volumes of data, which poses challenges for real-time processing and management. Edge computing provides a viable solution by enabling local data processing and reducing reliance on centralized data centers, thereby decreasing latency. However, data redundancy remains a key challenge that can negatively impact system performance. This study proposes an edge computing optimization framework that integrates K-Means clustering with decision tree-based data analytics. The proposed approach aims to reduce redundant data and improve data handling efficiency at the network edge. Simulation results indicate improvements in system performance compared to existing methods, including an increase in overall efficiency by 12%, an improvement in data analysis accuracy by 9%, a reduction in energy consumption by 11%, and a decrease in response time by 7%.